1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735 5736 5737 5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748 5749 5750 5751 5752 5753 5754 5755 5756 5757 5758 5759 5760 5761 5762 5763 5764 5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827 5828 5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843 5844 5845 5846 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866 5867 5868 5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000 6001 6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 6130 6131 6132 6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144 6145 6146 6147 6148 6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189
|
/*
** 2006 Oct 10
**
** The author disclaims copyright to this source code. In place of
** a legal notice, here is a blessing:
**
** May you do good and not evil.
** May you find forgiveness for yourself and forgive others.
** May you share freely, never taking more than you give.
**
******************************************************************************
**
** This is an SQLite module implementing full-text search.
*/
/*
** The code in this file is only compiled if:
**
** * The FTS3 module is being built as an extension
** (in which case SQLITE_CORE is not defined), or
**
** * The FTS3 module is being built into the core of
** SQLite (in which case SQLITE_ENABLE_FTS3 is defined).
*/
/* The full-text index is stored in a series of b+tree (-like)
** structures called segments which map terms to doclists. The
** structures are like b+trees in layout, but are constructed from the
** bottom up in optimal fashion and are not updatable. Since trees
** are built from the bottom up, things will be described from the
** bottom up.
**
**
**** Varints ****
** The basic unit of encoding is a variable-length integer called a
** varint. We encode variable-length integers in little-endian order
** using seven bits * per byte as follows:
**
** KEY:
** A = 0xxxxxxx 7 bits of data and one flag bit
** B = 1xxxxxxx 7 bits of data and one flag bit
**
** 7 bits - A
** 14 bits - BA
** 21 bits - BBA
** and so on.
**
** This is similar in concept to how sqlite encodes "varints" but
** the encoding is not the same. SQLite varints are big-endian
** are are limited to 9 bytes in length whereas FTS3 varints are
** little-endian and can be up to 10 bytes in length (in theory).
**
** Example encodings:
**
** 1: 0x01
** 127: 0x7f
** 128: 0x81 0x00
**
**
**** Document lists ****
** A doclist (document list) holds a docid-sorted list of hits for a
** given term. Doclists hold docids and associated token positions.
** A docid is the unique integer identifier for a single document.
** A position is the index of a word within the document. The first
** word of the document has a position of 0.
**
** FTS3 used to optionally store character offsets using a compile-time
** option. But that functionality is no longer supported.
**
** A doclist is stored like this:
**
** array {
** varint docid; (delta from previous doclist)
** array { (position list for column 0)
** varint position; (2 more than the delta from previous position)
** }
** array {
** varint POS_COLUMN; (marks start of position list for new column)
** varint column; (index of new column)
** array {
** varint position; (2 more than the delta from previous position)
** }
** }
** varint POS_END; (marks end of positions for this document.
** }
**
** Here, array { X } means zero or more occurrences of X, adjacent in
** memory. A "position" is an index of a token in the token stream
** generated by the tokenizer. Note that POS_END and POS_COLUMN occur
** in the same logical place as the position element, and act as sentinals
** ending a position list array. POS_END is 0. POS_COLUMN is 1.
** The positions numbers are not stored literally but rather as two more
** than the difference from the prior position, or the just the position plus
** 2 for the first position. Example:
**
** label: A B C D E F G H I J K
** value: 123 5 9 1 1 14 35 0 234 72 0
**
** The 123 value is the first docid. For column zero in this document
** there are two matches at positions 3 and 10 (5-2 and 9-2+3). The 1
** at D signals the start of a new column; the 1 at E indicates that the
** new column is column number 1. There are two positions at 12 and 45
** (14-2 and 35-2+12). The 0 at H indicate the end-of-document. The
** 234 at I is the delta to next docid (357). It has one position 70
** (72-2) and then terminates with the 0 at K.
**
** A "position-list" is the list of positions for multiple columns for
** a single docid. A "column-list" is the set of positions for a single
** column. Hence, a position-list consists of one or more column-lists,
** a document record consists of a docid followed by a position-list and
** a doclist consists of one or more document records.
**
** A bare doclist omits the position information, becoming an
** array of varint-encoded docids.
**
**** Segment leaf nodes ****
** Segment leaf nodes store terms and doclists, ordered by term. Leaf
** nodes are written using LeafWriter, and read using LeafReader (to
** iterate through a single leaf node's data) and LeavesReader (to
** iterate through a segment's entire leaf layer). Leaf nodes have
** the format:
**
** varint iHeight; (height from leaf level, always 0)
** varint nTerm; (length of first term)
** char pTerm[nTerm]; (content of first term)
** varint nDoclist; (length of term's associated doclist)
** char pDoclist[nDoclist]; (content of doclist)
** array {
** (further terms are delta-encoded)
** varint nPrefix; (length of prefix shared with previous term)
** varint nSuffix; (length of unshared suffix)
** char pTermSuffix[nSuffix];(unshared suffix of next term)
** varint nDoclist; (length of term's associated doclist)
** char pDoclist[nDoclist]; (content of doclist)
** }
**
** Here, array { X } means zero or more occurrences of X, adjacent in
** memory.
**
** Leaf nodes are broken into blocks which are stored contiguously in
** the %_segments table in sorted order. This means that when the end
** of a node is reached, the next term is in the node with the next
** greater node id.
**
** New data is spilled to a new leaf node when the current node
** exceeds LEAF_MAX bytes (default 2048). New data which itself is
** larger than STANDALONE_MIN (default 1024) is placed in a standalone
** node (a leaf node with a single term and doclist). The goal of
** these settings is to pack together groups of small doclists while
** making it efficient to directly access large doclists. The
** assumption is that large doclists represent terms which are more
** likely to be query targets.
**
** TODO(shess) It may be useful for blocking decisions to be more
** dynamic. For instance, it may make more sense to have a 2.5k leaf
** node rather than splitting into 2k and .5k nodes. My intuition is
** that this might extend through 2x or 4x the pagesize.
**
**
**** Segment interior nodes ****
** Segment interior nodes store blockids for subtree nodes and terms
** to describe what data is stored by the each subtree. Interior
** nodes are written using InteriorWriter, and read using
** InteriorReader. InteriorWriters are created as needed when
** SegmentWriter creates new leaf nodes, or when an interior node
** itself grows too big and must be split. The format of interior
** nodes:
**
** varint iHeight; (height from leaf level, always >0)
** varint iBlockid; (block id of node's leftmost subtree)
** optional {
** varint nTerm; (length of first term)
** char pTerm[nTerm]; (content of first term)
** array {
** (further terms are delta-encoded)
** varint nPrefix; (length of shared prefix with previous term)
** varint nSuffix; (length of unshared suffix)
** char pTermSuffix[nSuffix]; (unshared suffix of next term)
** }
** }
**
** Here, optional { X } means an optional element, while array { X }
** means zero or more occurrences of X, adjacent in memory.
**
** An interior node encodes n terms separating n+1 subtrees. The
** subtree blocks are contiguous, so only the first subtree's blockid
** is encoded. The subtree at iBlockid will contain all terms less
** than the first term encoded (or all terms if no term is encoded).
** Otherwise, for terms greater than or equal to pTerm[i] but less
** than pTerm[i+1], the subtree for that term will be rooted at
** iBlockid+i. Interior nodes only store enough term data to
** distinguish adjacent children (if the rightmost term of the left
** child is "something", and the leftmost term of the right child is
** "wicked", only "w" is stored).
**
** New data is spilled to a new interior node at the same height when
** the current node exceeds INTERIOR_MAX bytes (default 2048).
** INTERIOR_MIN_TERMS (default 7) keeps large terms from monopolizing
** interior nodes and making the tree too skinny. The interior nodes
** at a given height are naturally tracked by interior nodes at
** height+1, and so on.
**
**
**** Segment directory ****
** The segment directory in table %_segdir stores meta-information for
** merging and deleting segments, and also the root node of the
** segment's tree.
**
** The root node is the top node of the segment's tree after encoding
** the entire segment, restricted to ROOT_MAX bytes (default 1024).
** This could be either a leaf node or an interior node. If the top
** node requires more than ROOT_MAX bytes, it is flushed to %_segments
** and a new root interior node is generated (which should always fit
** within ROOT_MAX because it only needs space for 2 varints, the
** height and the blockid of the previous root).
**
** The meta-information in the segment directory is:
** level - segment level (see below)
** idx - index within level
** - (level,idx uniquely identify a segment)
** start_block - first leaf node
** leaves_end_block - last leaf node
** end_block - last block (including interior nodes)
** root - contents of root node
**
** If the root node is a leaf node, then start_block,
** leaves_end_block, and end_block are all 0.
**
**
**** Segment merging ****
** To amortize update costs, segments are grouped into levels and
** merged in batches. Each increase in level represents exponentially
** more documents.
**
** New documents (actually, document updates) are tokenized and
** written individually (using LeafWriter) to a level 0 segment, with
** incrementing idx. When idx reaches MERGE_COUNT (default 16), all
** level 0 segments are merged into a single level 1 segment. Level 1
** is populated like level 0, and eventually MERGE_COUNT level 1
** segments are merged to a single level 2 segment (representing
** MERGE_COUNT^2 updates), and so on.
**
** A segment merge traverses all segments at a given level in
** parallel, performing a straightforward sorted merge. Since segment
** leaf nodes are written in to the %_segments table in order, this
** merge traverses the underlying sqlite disk structures efficiently.
** After the merge, all segment blocks from the merged level are
** deleted.
**
** MERGE_COUNT controls how often we merge segments. 16 seems to be
** somewhat of a sweet spot for insertion performance. 32 and 64 show
** very similar performance numbers to 16 on insertion, though they're
** a tiny bit slower (perhaps due to more overhead in merge-time
** sorting). 8 is about 20% slower than 16, 4 about 50% slower than
** 16, 2 about 66% slower than 16.
**
** At query time, high MERGE_COUNT increases the number of segments
** which need to be scanned and merged. For instance, with 100k docs
** inserted:
**
** MERGE_COUNT segments
** 16 25
** 8 12
** 4 10
** 2 6
**
** This appears to have only a moderate impact on queries for very
** frequent terms (which are somewhat dominated by segment merge
** costs), and infrequent and non-existent terms still seem to be fast
** even with many segments.
**
** TODO(shess) That said, it would be nice to have a better query-side
** argument for MERGE_COUNT of 16. Also, it is possible/likely that
** optimizations to things like doclist merging will swing the sweet
** spot around.
**
**
**
**** Handling of deletions and updates ****
** Since we're using a segmented structure, with no docid-oriented
** index into the term index, we clearly cannot simply update the term
** index when a document is deleted or updated. For deletions, we
** write an empty doclist (varint(docid) varint(POS_END)), for updates
** we simply write the new doclist. Segment merges overwrite older
** data for a particular docid with newer data, so deletes or updates
** will eventually overtake the earlier data and knock it out. The
** query logic likewise merges doclists so that newer data knocks out
** older data.
*/
#include "fts3Int.h"
#if !defined(SQLITE_CORE) || defined(SQLITE_ENABLE_FTS3)
#if defined(SQLITE_ENABLE_FTS3) && !defined(SQLITE_CORE)
# define SQLITE_CORE 1
#endif
#include <assert.h>
#include <stdlib.h>
#include <stddef.h>
#include <stdio.h>
#include <string.h>
#include <stdarg.h>
#include "fts3.h"
#ifndef SQLITE_CORE
# include "sqlite3ext.h"
SQLITE_EXTENSION_INIT1
#endif
typedef struct Fts3HashWrapper Fts3HashWrapper;
struct Fts3HashWrapper {
Fts3Hash hash; /* Hash table */
int nRef; /* Number of pointers to this object */
};
static int fts3EvalNext(Fts3Cursor *pCsr);
static int fts3EvalStart(Fts3Cursor *pCsr);
static int fts3TermSegReaderCursor(
Fts3Cursor *, const char *, int, int, Fts3MultiSegReader **);
/*
** This variable is set to false when running tests for which the on disk
** structures should not be corrupt. Otherwise, true. If it is false, extra
** assert() conditions in the fts3 code are activated - conditions that are
** only true if it is guaranteed that the fts3 database is not corrupt.
*/
#ifdef SQLITE_DEBUG
int sqlite3_fts3_may_be_corrupt = 1;
#endif
/*
** Write a 64-bit variable-length integer to memory starting at p[0].
** The length of data written will be between 1 and FTS3_VARINT_MAX bytes.
** The number of bytes written is returned.
*/
int sqlite3Fts3PutVarint(char *p, sqlite_int64 v){
unsigned char *q = (unsigned char *) p;
sqlite_uint64 vu = v;
do{
*q++ = (unsigned char) ((vu & 0x7f) | 0x80);
vu >>= 7;
}while( vu!=0 );
q[-1] &= 0x7f; /* turn off high bit in final byte */
assert( q - (unsigned char *)p <= FTS3_VARINT_MAX );
return (int) (q - (unsigned char *)p);
}
#define GETVARINT_STEP(v, ptr, shift, mask1, mask2, var, ret) \
v = (v & mask1) | ( (*(const unsigned char*)(ptr++)) << shift ); \
if( (v & mask2)==0 ){ var = v; return ret; }
#define GETVARINT_INIT(v, ptr, shift, mask1, mask2, var, ret) \
v = (*ptr++); \
if( (v & mask2)==0 ){ var = v; return ret; }
int sqlite3Fts3GetVarintU(const char *pBuf, sqlite_uint64 *v){
const unsigned char *p = (const unsigned char*)pBuf;
const unsigned char *pStart = p;
u32 a;
u64 b;
int shift;
GETVARINT_INIT(a, p, 0, 0x00, 0x80, *v, 1);
GETVARINT_STEP(a, p, 7, 0x7F, 0x4000, *v, 2);
GETVARINT_STEP(a, p, 14, 0x3FFF, 0x200000, *v, 3);
GETVARINT_STEP(a, p, 21, 0x1FFFFF, 0x10000000, *v, 4);
b = (a & 0x0FFFFFFF );
for(shift=28; shift<=63; shift+=7){
u64 c = *p++;
b += (c&0x7F) << shift;
if( (c & 0x80)==0 ) break;
}
*v = b;
return (int)(p - pStart);
}
/*
** Read a 64-bit variable-length integer from memory starting at p[0].
** Return the number of bytes read, or 0 on error.
** The value is stored in *v.
*/
int sqlite3Fts3GetVarint(const char *pBuf, sqlite_int64 *v){
return sqlite3Fts3GetVarintU(pBuf, (sqlite3_uint64*)v);
}
/*
** Read a 64-bit variable-length integer from memory starting at p[0] and
** not extending past pEnd[-1].
** Return the number of bytes read, or 0 on error.
** The value is stored in *v.
*/
int sqlite3Fts3GetVarintBounded(
const char *pBuf,
const char *pEnd,
sqlite_int64 *v
){
const unsigned char *p = (const unsigned char*)pBuf;
const unsigned char *pStart = p;
const unsigned char *pX = (const unsigned char*)pEnd;
u64 b = 0;
int shift;
for(shift=0; shift<=63; shift+=7){
u64 c = p<pX ? *p : 0;
p++;
b += (c&0x7F) << shift;
if( (c & 0x80)==0 ) break;
}
*v = b;
return (int)(p - pStart);
}
/*
** Similar to sqlite3Fts3GetVarint(), except that the output is truncated to
** a non-negative 32-bit integer before it is returned.
*/
int sqlite3Fts3GetVarint32(const char *p, int *pi){
const unsigned char *ptr = (const unsigned char*)p;
u32 a;
#ifndef fts3GetVarint32
GETVARINT_INIT(a, ptr, 0, 0x00, 0x80, *pi, 1);
#else
a = (*ptr++);
assert( a & 0x80 );
#endif
GETVARINT_STEP(a, ptr, 7, 0x7F, 0x4000, *pi, 2);
GETVARINT_STEP(a, ptr, 14, 0x3FFF, 0x200000, *pi, 3);
GETVARINT_STEP(a, ptr, 21, 0x1FFFFF, 0x10000000, *pi, 4);
a = (a & 0x0FFFFFFF );
*pi = (int)(a | ((u32)(*ptr & 0x07) << 28));
assert( 0==(a & 0x80000000) );
assert( *pi>=0 );
return 5;
}
/*
** Return the number of bytes required to encode v as a varint
*/
int sqlite3Fts3VarintLen(sqlite3_uint64 v){
int i = 0;
do{
i++;
v >>= 7;
}while( v!=0 );
return i;
}
/*
** Convert an SQL-style quoted string into a normal string by removing
** the quote characters. The conversion is done in-place. If the
** input does not begin with a quote character, then this routine
** is a no-op.
**
** Examples:
**
** "abc" becomes abc
** 'xyz' becomes xyz
** [pqr] becomes pqr
** `mno` becomes mno
**
*/
void sqlite3Fts3Dequote(char *z){
char quote; /* Quote character (if any ) */
quote = z[0];
if( quote=='[' || quote=='\'' || quote=='"' || quote=='`' ){
int iIn = 1; /* Index of next byte to read from input */
int iOut = 0; /* Index of next byte to write to output */
/* If the first byte was a '[', then the close-quote character is a ']' */
if( quote=='[' ) quote = ']';
while( z[iIn] ){
if( z[iIn]==quote ){
if( z[iIn+1]!=quote ) break;
z[iOut++] = quote;
iIn += 2;
}else{
z[iOut++] = z[iIn++];
}
}
z[iOut] = '\0';
}
}
/*
** Read a single varint from the doclist at *pp and advance *pp to point
** to the first byte past the end of the varint. Add the value of the varint
** to *pVal.
*/
static void fts3GetDeltaVarint(char **pp, sqlite3_int64 *pVal){
sqlite3_int64 iVal;
*pp += sqlite3Fts3GetVarint(*pp, &iVal);
*pVal += iVal;
}
/*
** When this function is called, *pp points to the first byte following a
** varint that is part of a doclist (or position-list, or any other list
** of varints). This function moves *pp to point to the start of that varint,
** and sets *pVal by the varint value.
**
** Argument pStart points to the first byte of the doclist that the
** varint is part of.
*/
static void fts3GetReverseVarint(
char **pp,
char *pStart,
sqlite3_int64 *pVal
){
sqlite3_int64 iVal;
char *p;
/* Pointer p now points at the first byte past the varint we are
** interested in. So, unless the doclist is corrupt, the 0x80 bit is
** clear on character p[-1]. */
for(p = (*pp)-2; p>=pStart && *p&0x80; p--);
p++;
*pp = p;
sqlite3Fts3GetVarint(p, &iVal);
*pVal = iVal;
}
/*
** The xDisconnect() virtual table method.
*/
static int fts3DisconnectMethod(sqlite3_vtab *pVtab){
Fts3Table *p = (Fts3Table *)pVtab;
int i;
assert( p->nPendingData==0 );
assert( p->pSegments==0 );
/* Free any prepared statements held */
sqlite3_finalize(p->pSeekStmt);
for(i=0; i<SizeofArray(p->aStmt); i++){
sqlite3_finalize(p->aStmt[i]);
}
sqlite3_free(p->zSegmentsTbl);
sqlite3_free(p->zReadExprlist);
sqlite3_free(p->zWriteExprlist);
sqlite3_free(p->zContentTbl);
sqlite3_free(p->zLanguageid);
/* Invoke the tokenizer destructor to free the tokenizer. */
p->pTokenizer->pModule->xDestroy(p->pTokenizer);
sqlite3_free(p);
return SQLITE_OK;
}
/*
** Write an error message into *pzErr
*/
void sqlite3Fts3ErrMsg(char **pzErr, const char *zFormat, ...){
va_list ap;
sqlite3_free(*pzErr);
va_start(ap, zFormat);
*pzErr = sqlite3_vmprintf(zFormat, ap);
va_end(ap);
}
/*
** Construct one or more SQL statements from the format string given
** and then evaluate those statements. The success code is written
** into *pRc.
**
** If *pRc is initially non-zero then this routine is a no-op.
*/
static void fts3DbExec(
int *pRc, /* Success code */
sqlite3 *db, /* Database in which to run SQL */
const char *zFormat, /* Format string for SQL */
... /* Arguments to the format string */
){
va_list ap;
char *zSql;
if( *pRc ) return;
va_start(ap, zFormat);
zSql = sqlite3_vmprintf(zFormat, ap);
va_end(ap);
if( zSql==0 ){
*pRc = SQLITE_NOMEM;
}else{
*pRc = sqlite3_exec(db, zSql, 0, 0, 0);
sqlite3_free(zSql);
}
}
/*
** The xDestroy() virtual table method.
*/
static int fts3DestroyMethod(sqlite3_vtab *pVtab){
Fts3Table *p = (Fts3Table *)pVtab;
int rc = SQLITE_OK; /* Return code */
const char *zDb = p->zDb; /* Name of database (e.g. "main", "temp") */
sqlite3 *db = p->db; /* Database handle */
/* Drop the shadow tables */
fts3DbExec(&rc, db,
"DROP TABLE IF EXISTS %Q.'%q_segments';"
"DROP TABLE IF EXISTS %Q.'%q_segdir';"
"DROP TABLE IF EXISTS %Q.'%q_docsize';"
"DROP TABLE IF EXISTS %Q.'%q_stat';"
"%s DROP TABLE IF EXISTS %Q.'%q_content';",
zDb, p->zName,
zDb, p->zName,
zDb, p->zName,
zDb, p->zName,
(p->zContentTbl ? "--" : ""), zDb,p->zName
);
/* If everything has worked, invoke fts3DisconnectMethod() to free the
** memory associated with the Fts3Table structure and return SQLITE_OK.
** Otherwise, return an SQLite error code.
*/
return (rc==SQLITE_OK ? fts3DisconnectMethod(pVtab) : rc);
}
/*
** Invoke sqlite3_declare_vtab() to declare the schema for the FTS3 table
** passed as the first argument. This is done as part of the xConnect()
** and xCreate() methods.
**
** If *pRc is non-zero when this function is called, it is a no-op.
** Otherwise, if an error occurs, an SQLite error code is stored in *pRc
** before returning.
*/
static void fts3DeclareVtab(int *pRc, Fts3Table *p){
if( *pRc==SQLITE_OK ){
int i; /* Iterator variable */
int rc; /* Return code */
char *zSql; /* SQL statement passed to declare_vtab() */
char *zCols; /* List of user defined columns */
const char *zLanguageid;
zLanguageid = (p->zLanguageid ? p->zLanguageid : "__langid");
sqlite3_vtab_config(p->db, SQLITE_VTAB_CONSTRAINT_SUPPORT, 1);
sqlite3_vtab_config(p->db, SQLITE_VTAB_INNOCUOUS);
/* Create a list of user columns for the virtual table */
zCols = sqlite3_mprintf("%Q, ", p->azColumn[0]);
for(i=1; zCols && i<p->nColumn; i++){
zCols = sqlite3_mprintf("%z%Q, ", zCols, p->azColumn[i]);
}
/* Create the whole "CREATE TABLE" statement to pass to SQLite */
zSql = sqlite3_mprintf(
"CREATE TABLE x(%s %Q HIDDEN, docid HIDDEN, %Q HIDDEN)",
zCols, p->zName, zLanguageid
);
if( !zCols || !zSql ){
rc = SQLITE_NOMEM;
}else{
rc = sqlite3_declare_vtab(p->db, zSql);
}
sqlite3_free(zSql);
sqlite3_free(zCols);
*pRc = rc;
}
}
/*
** Create the %_stat table if it does not already exist.
*/
void sqlite3Fts3CreateStatTable(int *pRc, Fts3Table *p){
fts3DbExec(pRc, p->db,
"CREATE TABLE IF NOT EXISTS %Q.'%q_stat'"
"(id INTEGER PRIMARY KEY, value BLOB);",
p->zDb, p->zName
);
if( (*pRc)==SQLITE_OK ) p->bHasStat = 1;
}
/*
** Create the backing store tables (%_content, %_segments and %_segdir)
** required by the FTS3 table passed as the only argument. This is done
** as part of the vtab xCreate() method.
**
** If the p->bHasDocsize boolean is true (indicating that this is an
** FTS4 table, not an FTS3 table) then also create the %_docsize and
** %_stat tables required by FTS4.
*/
static int fts3CreateTables(Fts3Table *p){
int rc = SQLITE_OK; /* Return code */
int i; /* Iterator variable */
sqlite3 *db = p->db; /* The database connection */
if( p->zContentTbl==0 ){
const char *zLanguageid = p->zLanguageid;
char *zContentCols; /* Columns of %_content table */
/* Create a list of user columns for the content table */
zContentCols = sqlite3_mprintf("docid INTEGER PRIMARY KEY");
for(i=0; zContentCols && i<p->nColumn; i++){
char *z = p->azColumn[i];
zContentCols = sqlite3_mprintf("%z, 'c%d%q'", zContentCols, i, z);
}
if( zLanguageid && zContentCols ){
zContentCols = sqlite3_mprintf("%z, langid", zContentCols, zLanguageid);
}
if( zContentCols==0 ) rc = SQLITE_NOMEM;
/* Create the content table */
fts3DbExec(&rc, db,
"CREATE TABLE %Q.'%q_content'(%s)",
p->zDb, p->zName, zContentCols
);
sqlite3_free(zContentCols);
}
/* Create other tables */
fts3DbExec(&rc, db,
"CREATE TABLE %Q.'%q_segments'(blockid INTEGER PRIMARY KEY, block BLOB);",
p->zDb, p->zName
);
fts3DbExec(&rc, db,
"CREATE TABLE %Q.'%q_segdir'("
"level INTEGER,"
"idx INTEGER,"
"start_block INTEGER,"
"leaves_end_block INTEGER,"
"end_block INTEGER,"
"root BLOB,"
"PRIMARY KEY(level, idx)"
");",
p->zDb, p->zName
);
if( p->bHasDocsize ){
fts3DbExec(&rc, db,
"CREATE TABLE %Q.'%q_docsize'(docid INTEGER PRIMARY KEY, size BLOB);",
p->zDb, p->zName
);
}
assert( p->bHasStat==p->bFts4 );
if( p->bHasStat ){
sqlite3Fts3CreateStatTable(&rc, p);
}
return rc;
}
/*
** Store the current database page-size in bytes in p->nPgsz.
**
** If *pRc is non-zero when this function is called, it is a no-op.
** Otherwise, if an error occurs, an SQLite error code is stored in *pRc
** before returning.
*/
static void fts3DatabasePageSize(int *pRc, Fts3Table *p){
if( *pRc==SQLITE_OK ){
int rc; /* Return code */
char *zSql; /* SQL text "PRAGMA %Q.page_size" */
sqlite3_stmt *pStmt; /* Compiled "PRAGMA %Q.page_size" statement */
zSql = sqlite3_mprintf("PRAGMA %Q.page_size", p->zDb);
if( !zSql ){
rc = SQLITE_NOMEM;
}else{
rc = sqlite3_prepare(p->db, zSql, -1, &pStmt, 0);
if( rc==SQLITE_OK ){
sqlite3_step(pStmt);
p->nPgsz = sqlite3_column_int(pStmt, 0);
rc = sqlite3_finalize(pStmt);
}else if( rc==SQLITE_AUTH ){
p->nPgsz = 1024;
rc = SQLITE_OK;
}
}
assert( p->nPgsz>0 || rc!=SQLITE_OK );
sqlite3_free(zSql);
*pRc = rc;
}
}
/*
** "Special" FTS4 arguments are column specifications of the following form:
**
** <key> = <value>
**
** There may not be whitespace surrounding the "=" character. The <value>
** term may be quoted, but the <key> may not.
*/
static int fts3IsSpecialColumn(
const char *z,
int *pnKey,
char **pzValue
){
char *zValue;
const char *zCsr = z;
while( *zCsr!='=' ){
if( *zCsr=='\0' ) return 0;
zCsr++;
}
*pnKey = (int)(zCsr-z);
zValue = sqlite3_mprintf("%s", &zCsr[1]);
if( zValue ){
sqlite3Fts3Dequote(zValue);
}
*pzValue = zValue;
return 1;
}
/*
** Append the output of a printf() style formatting to an existing string.
*/
static void fts3Appendf(
int *pRc, /* IN/OUT: Error code */
char **pz, /* IN/OUT: Pointer to string buffer */
const char *zFormat, /* Printf format string to append */
... /* Arguments for printf format string */
){
if( *pRc==SQLITE_OK ){
va_list ap;
char *z;
va_start(ap, zFormat);
z = sqlite3_vmprintf(zFormat, ap);
va_end(ap);
if( z && *pz ){
char *z2 = sqlite3_mprintf("%s%s", *pz, z);
sqlite3_free(z);
z = z2;
}
if( z==0 ) *pRc = SQLITE_NOMEM;
sqlite3_free(*pz);
*pz = z;
}
}
/*
** Return a copy of input string zInput enclosed in double-quotes (") and
** with all double quote characters escaped. For example:
**
** fts3QuoteId("un \"zip\"") -> "un \"\"zip\"\""
**
** The pointer returned points to memory obtained from sqlite3_malloc(). It
** is the callers responsibility to call sqlite3_free() to release this
** memory.
*/
static char *fts3QuoteId(char const *zInput){
sqlite3_int64 nRet;
char *zRet;
nRet = 2 + (int)strlen(zInput)*2 + 1;
zRet = sqlite3_malloc64(nRet);
if( zRet ){
int i;
char *z = zRet;
*(z++) = '"';
for(i=0; zInput[i]; i++){
if( zInput[i]=='"' ) *(z++) = '"';
*(z++) = zInput[i];
}
*(z++) = '"';
*(z++) = '\0';
}
return zRet;
}
/*
** Return a list of comma separated SQL expressions and a FROM clause that
** could be used in a SELECT statement such as the following:
**
** SELECT <list of expressions> FROM %_content AS x ...
**
** to return the docid, followed by each column of text data in order
** from left to write. If parameter zFunc is not NULL, then instead of
** being returned directly each column of text data is passed to an SQL
** function named zFunc first. For example, if zFunc is "unzip" and the
** table has the three user-defined columns "a", "b", and "c", the following
** string is returned:
**
** "docid, unzip(x.'a'), unzip(x.'b'), unzip(x.'c') FROM %_content AS x"
**
** The pointer returned points to a buffer allocated by sqlite3_malloc(). It
** is the responsibility of the caller to eventually free it.
**
** If *pRc is not SQLITE_OK when this function is called, it is a no-op (and
** a NULL pointer is returned). Otherwise, if an OOM error is encountered
** by this function, NULL is returned and *pRc is set to SQLITE_NOMEM. If
** no error occurs, *pRc is left unmodified.
*/
static char *fts3ReadExprList(Fts3Table *p, const char *zFunc, int *pRc){
char *zRet = 0;
char *zFree = 0;
char *zFunction;
int i;
if( p->zContentTbl==0 ){
if( !zFunc ){
zFunction = "";
}else{
zFree = zFunction = fts3QuoteId(zFunc);
}
fts3Appendf(pRc, &zRet, "docid");
for(i=0; i<p->nColumn; i++){
fts3Appendf(pRc, &zRet, ",%s(x.'c%d%q')", zFunction, i, p->azColumn[i]);
}
if( p->zLanguageid ){
fts3Appendf(pRc, &zRet, ", x.%Q", "langid");
}
sqlite3_free(zFree);
}else{
fts3Appendf(pRc, &zRet, "rowid");
for(i=0; i<p->nColumn; i++){
fts3Appendf(pRc, &zRet, ", x.'%q'", p->azColumn[i]);
}
if( p->zLanguageid ){
fts3Appendf(pRc, &zRet, ", x.%Q", p->zLanguageid);
}
}
fts3Appendf(pRc, &zRet, " FROM '%q'.'%q%s' AS x",
p->zDb,
(p->zContentTbl ? p->zContentTbl : p->zName),
(p->zContentTbl ? "" : "_content")
);
return zRet;
}
/*
** Return a list of N comma separated question marks, where N is the number
** of columns in the %_content table (one for the docid plus one for each
** user-defined text column).
**
** If argument zFunc is not NULL, then all but the first question mark
** is preceded by zFunc and an open bracket, and followed by a closed
** bracket. For example, if zFunc is "zip" and the FTS3 table has three
** user-defined text columns, the following string is returned:
**
** "?, zip(?), zip(?), zip(?)"
**
** The pointer returned points to a buffer allocated by sqlite3_malloc(). It
** is the responsibility of the caller to eventually free it.
**
** If *pRc is not SQLITE_OK when this function is called, it is a no-op (and
** a NULL pointer is returned). Otherwise, if an OOM error is encountered
** by this function, NULL is returned and *pRc is set to SQLITE_NOMEM. If
** no error occurs, *pRc is left unmodified.
*/
static char *fts3WriteExprList(Fts3Table *p, const char *zFunc, int *pRc){
char *zRet = 0;
char *zFree = 0;
char *zFunction;
int i;
if( !zFunc ){
zFunction = "";
}else{
zFree = zFunction = fts3QuoteId(zFunc);
}
fts3Appendf(pRc, &zRet, "?");
for(i=0; i<p->nColumn; i++){
fts3Appendf(pRc, &zRet, ",%s(?)", zFunction);
}
if( p->zLanguageid ){
fts3Appendf(pRc, &zRet, ", ?");
}
sqlite3_free(zFree);
return zRet;
}
/*
** Buffer z contains a positive integer value encoded as utf-8 text.
** Decode this value and store it in *pnOut, returning the number of bytes
** consumed. If an overflow error occurs return a negative value.
*/
int sqlite3Fts3ReadInt(const char *z, int *pnOut){
u64 iVal = 0;
int i;
for(i=0; z[i]>='0' && z[i]<='9'; i++){
iVal = iVal*10 + (z[i] - '0');
if( iVal>0x7FFFFFFF ) return -1;
}
*pnOut = (int)iVal;
return i;
}
/*
** This function interprets the string at (*pp) as a non-negative integer
** value. It reads the integer and sets *pnOut to the value read, then
** sets *pp to point to the byte immediately following the last byte of
** the integer value.
**
** Only decimal digits ('0'..'9') may be part of an integer value.
**
** If *pp does not being with a decimal digit SQLITE_ERROR is returned and
** the output value undefined. Otherwise SQLITE_OK is returned.
**
** This function is used when parsing the "prefix=" FTS4 parameter.
*/
static int fts3GobbleInt(const char **pp, int *pnOut){
const int MAX_NPREFIX = 10000000;
int nInt = 0; /* Output value */
int nByte;
nByte = sqlite3Fts3ReadInt(*pp, &nInt);
if( nInt>MAX_NPREFIX ){
nInt = 0;
}
if( nByte==0 ){
return SQLITE_ERROR;
}
*pnOut = nInt;
*pp += nByte;
return SQLITE_OK;
}
/*
** This function is called to allocate an array of Fts3Index structures
** representing the indexes maintained by the current FTS table. FTS tables
** always maintain the main "terms" index, but may also maintain one or
** more "prefix" indexes, depending on the value of the "prefix=" parameter
** (if any) specified as part of the CREATE VIRTUAL TABLE statement.
**
** Argument zParam is passed the value of the "prefix=" option if one was
** specified, or NULL otherwise.
**
** If no error occurs, SQLITE_OK is returned and *apIndex set to point to
** the allocated array. *pnIndex is set to the number of elements in the
** array. If an error does occur, an SQLite error code is returned.
**
** Regardless of whether or not an error is returned, it is the responsibility
** of the caller to call sqlite3_free() on the output array to free it.
*/
static int fts3PrefixParameter(
const char *zParam, /* ABC in prefix=ABC parameter to parse */
int *pnIndex, /* OUT: size of *apIndex[] array */
struct Fts3Index **apIndex /* OUT: Array of indexes for this table */
){
struct Fts3Index *aIndex; /* Allocated array */
int nIndex = 1; /* Number of entries in array */
if( zParam && zParam[0] ){
const char *p;
nIndex++;
for(p=zParam; *p; p++){
if( *p==',' ) nIndex++;
}
}
aIndex = sqlite3_malloc64(sizeof(struct Fts3Index) * nIndex);
*apIndex = aIndex;
if( !aIndex ){
return SQLITE_NOMEM;
}
memset(aIndex, 0, sizeof(struct Fts3Index) * nIndex);
if( zParam ){
const char *p = zParam;
int i;
for(i=1; i<nIndex; i++){
int nPrefix = 0;
if( fts3GobbleInt(&p, &nPrefix) ) return SQLITE_ERROR;
assert( nPrefix>=0 );
if( nPrefix==0 ){
nIndex--;
i--;
}else{
aIndex[i].nPrefix = nPrefix;
}
p++;
}
}
*pnIndex = nIndex;
return SQLITE_OK;
}
/*
** This function is called when initializing an FTS4 table that uses the
** content=xxx option. It determines the number of and names of the columns
** of the new FTS4 table.
**
** The third argument passed to this function is the value passed to the
** config=xxx option (i.e. "xxx"). This function queries the database for
** a table of that name. If found, the output variables are populated
** as follows:
**
** *pnCol: Set to the number of columns table xxx has,
**
** *pnStr: Set to the total amount of space required to store a copy
** of each columns name, including the nul-terminator.
**
** *pazCol: Set to point to an array of *pnCol strings. Each string is
** the name of the corresponding column in table xxx. The array
** and its contents are allocated using a single allocation. It
** is the responsibility of the caller to free this allocation
** by eventually passing the *pazCol value to sqlite3_free().
**
** If the table cannot be found, an error code is returned and the output
** variables are undefined. Or, if an OOM is encountered, SQLITE_NOMEM is
** returned (and the output variables are undefined).
*/
static int fts3ContentColumns(
sqlite3 *db, /* Database handle */
const char *zDb, /* Name of db (i.e. "main", "temp" etc.) */
const char *zTbl, /* Name of content table */
const char ***pazCol, /* OUT: Malloc'd array of column names */
int *pnCol, /* OUT: Size of array *pazCol */
int *pnStr, /* OUT: Bytes of string content */
char **pzErr /* OUT: error message */
){
int rc = SQLITE_OK; /* Return code */
char *zSql; /* "SELECT *" statement on zTbl */
sqlite3_stmt *pStmt = 0; /* Compiled version of zSql */
zSql = sqlite3_mprintf("SELECT * FROM %Q.%Q", zDb, zTbl);
if( !zSql ){
rc = SQLITE_NOMEM;
}else{
rc = sqlite3_prepare(db, zSql, -1, &pStmt, 0);
if( rc!=SQLITE_OK ){
sqlite3Fts3ErrMsg(pzErr, "%s", sqlite3_errmsg(db));
}
}
sqlite3_free(zSql);
if( rc==SQLITE_OK ){
const char **azCol; /* Output array */
sqlite3_int64 nStr = 0; /* Size of all column names (incl. 0x00) */
int nCol; /* Number of table columns */
int i; /* Used to iterate through columns */
/* Loop through the returned columns. Set nStr to the number of bytes of
** space required to store a copy of each column name, including the
** nul-terminator byte. */
nCol = sqlite3_column_count(pStmt);
for(i=0; i<nCol; i++){
const char *zCol = sqlite3_column_name(pStmt, i);
nStr += strlen(zCol) + 1;
}
/* Allocate and populate the array to return. */
azCol = (const char **)sqlite3_malloc64(sizeof(char *) * nCol + nStr);
if( azCol==0 ){
rc = SQLITE_NOMEM;
}else{
char *p = (char *)&azCol[nCol];
for(i=0; i<nCol; i++){
const char *zCol = sqlite3_column_name(pStmt, i);
int n = (int)strlen(zCol)+1;
memcpy(p, zCol, n);
azCol[i] = p;
p += n;
}
}
sqlite3_finalize(pStmt);
/* Set the output variables. */
*pnCol = nCol;
*pnStr = nStr;
*pazCol = azCol;
}
return rc;
}
/*
** This function is the implementation of both the xConnect and xCreate
** methods of the FTS3 virtual table.
**
** The argv[] array contains the following:
**
** argv[0] -> module name ("fts3" or "fts4")
** argv[1] -> database name
** argv[2] -> table name
** argv[...] -> "column name" and other module argument fields.
*/
static int fts3InitVtab(
int isCreate, /* True for xCreate, false for xConnect */
sqlite3 *db, /* The SQLite database connection */
void *pAux, /* Hash table containing tokenizers */
int argc, /* Number of elements in argv array */
const char * const *argv, /* xCreate/xConnect argument array */
sqlite3_vtab **ppVTab, /* Write the resulting vtab structure here */
char **pzErr /* Write any error message here */
){
Fts3Hash *pHash = &((Fts3HashWrapper*)pAux)->hash;
Fts3Table *p = 0; /* Pointer to allocated vtab */
int rc = SQLITE_OK; /* Return code */
int i; /* Iterator variable */
sqlite3_int64 nByte; /* Size of allocation used for *p */
int iCol; /* Column index */
int nString = 0; /* Bytes required to hold all column names */
int nCol = 0; /* Number of columns in the FTS table */
char *zCsr; /* Space for holding column names */
int nDb; /* Bytes required to hold database name */
int nName; /* Bytes required to hold table name */
int isFts4 = (argv[0][3]=='4'); /* True for FTS4, false for FTS3 */
const char **aCol; /* Array of column names */
sqlite3_tokenizer *pTokenizer = 0; /* Tokenizer for this table */
int nIndex = 0; /* Size of aIndex[] array */
struct Fts3Index *aIndex = 0; /* Array of indexes for this table */
/* The results of parsing supported FTS4 key=value options: */
int bNoDocsize = 0; /* True to omit %_docsize table */
int bDescIdx = 0; /* True to store descending indexes */
char *zPrefix = 0; /* Prefix parameter value (or NULL) */
char *zCompress = 0; /* compress=? parameter (or NULL) */
char *zUncompress = 0; /* uncompress=? parameter (or NULL) */
char *zContent = 0; /* content=? parameter (or NULL) */
char *zLanguageid = 0; /* languageid=? parameter (or NULL) */
char **azNotindexed = 0; /* The set of notindexed= columns */
int nNotindexed = 0; /* Size of azNotindexed[] array */
assert( strlen(argv[0])==4 );
assert( (sqlite3_strnicmp(argv[0], "fts4", 4)==0 && isFts4)
|| (sqlite3_strnicmp(argv[0], "fts3", 4)==0 && !isFts4)
);
nDb = (int)strlen(argv[1]) + 1;
nName = (int)strlen(argv[2]) + 1;
nByte = sizeof(const char *) * (argc-2);
aCol = (const char **)sqlite3_malloc64(nByte);
if( aCol ){
memset((void*)aCol, 0, nByte);
azNotindexed = (char **)sqlite3_malloc64(nByte);
}
if( azNotindexed ){
memset(azNotindexed, 0, nByte);
}
if( !aCol || !azNotindexed ){
rc = SQLITE_NOMEM;
goto fts3_init_out;
}
/* Loop through all of the arguments passed by the user to the FTS3/4
** module (i.e. all the column names and special arguments). This loop
** does the following:
**
** + Figures out the number of columns the FTSX table will have, and
** the number of bytes of space that must be allocated to store copies
** of the column names.
**
** + If there is a tokenizer specification included in the arguments,
** initializes the tokenizer pTokenizer.
*/
for(i=3; rc==SQLITE_OK && i<argc; i++){
char const *z = argv[i];
int nKey;
char *zVal;
/* Check if this is a tokenizer specification */
if( !pTokenizer
&& strlen(z)>8
&& 0==sqlite3_strnicmp(z, "tokenize", 8)
&& 0==sqlite3Fts3IsIdChar(z[8])
){
rc = sqlite3Fts3InitTokenizer(pHash, &z[9], &pTokenizer, pzErr);
}
/* Check if it is an FTS4 special argument. */
else if( isFts4 && fts3IsSpecialColumn(z, &nKey, &zVal) ){
struct Fts4Option {
const char *zOpt;
int nOpt;
} aFts4Opt[] = {
{ "matchinfo", 9 }, /* 0 -> MATCHINFO */
{ "prefix", 6 }, /* 1 -> PREFIX */
{ "compress", 8 }, /* 2 -> COMPRESS */
{ "uncompress", 10 }, /* 3 -> UNCOMPRESS */
{ "order", 5 }, /* 4 -> ORDER */
{ "content", 7 }, /* 5 -> CONTENT */
{ "languageid", 10 }, /* 6 -> LANGUAGEID */
{ "notindexed", 10 } /* 7 -> NOTINDEXED */
};
int iOpt;
if( !zVal ){
rc = SQLITE_NOMEM;
}else{
for(iOpt=0; iOpt<SizeofArray(aFts4Opt); iOpt++){
struct Fts4Option *pOp = &aFts4Opt[iOpt];
if( nKey==pOp->nOpt && !sqlite3_strnicmp(z, pOp->zOpt, pOp->nOpt) ){
break;
}
}
switch( iOpt ){
case 0: /* MATCHINFO */
if( strlen(zVal)!=4 || sqlite3_strnicmp(zVal, "fts3", 4) ){
sqlite3Fts3ErrMsg(pzErr, "unrecognized matchinfo: %s", zVal);
rc = SQLITE_ERROR;
}
bNoDocsize = 1;
break;
case 1: /* PREFIX */
sqlite3_free(zPrefix);
zPrefix = zVal;
zVal = 0;
break;
case 2: /* COMPRESS */
sqlite3_free(zCompress);
zCompress = zVal;
zVal = 0;
break;
case 3: /* UNCOMPRESS */
sqlite3_free(zUncompress);
zUncompress = zVal;
zVal = 0;
break;
case 4: /* ORDER */
if( (strlen(zVal)!=3 || sqlite3_strnicmp(zVal, "asc", 3))
&& (strlen(zVal)!=4 || sqlite3_strnicmp(zVal, "desc", 4))
){
sqlite3Fts3ErrMsg(pzErr, "unrecognized order: %s", zVal);
rc = SQLITE_ERROR;
}
bDescIdx = (zVal[0]=='d' || zVal[0]=='D');
break;
case 5: /* CONTENT */
sqlite3_free(zContent);
zContent = zVal;
zVal = 0;
break;
case 6: /* LANGUAGEID */
assert( iOpt==6 );
sqlite3_free(zLanguageid);
zLanguageid = zVal;
zVal = 0;
break;
case 7: /* NOTINDEXED */
azNotindexed[nNotindexed++] = zVal;
zVal = 0;
break;
default:
assert( iOpt==SizeofArray(aFts4Opt) );
sqlite3Fts3ErrMsg(pzErr, "unrecognized parameter: %s", z);
rc = SQLITE_ERROR;
break;
}
sqlite3_free(zVal);
}
}
/* Otherwise, the argument is a column name. */
else {
nString += (int)(strlen(z) + 1);
aCol[nCol++] = z;
}
}
/* If a content=xxx option was specified, the following:
**
** 1. Ignore any compress= and uncompress= options.
**
** 2. If no column names were specified as part of the CREATE VIRTUAL
** TABLE statement, use all columns from the content table.
*/
if( rc==SQLITE_OK && zContent ){
sqlite3_free(zCompress);
sqlite3_free(zUncompress);
zCompress = 0;
zUncompress = 0;
if( nCol==0 ){
sqlite3_free((void*)aCol);
aCol = 0;
rc = fts3ContentColumns(db, argv[1], zContent,&aCol,&nCol,&nString,pzErr);
/* If a languageid= option was specified, remove the language id
** column from the aCol[] array. */
if( rc==SQLITE_OK && zLanguageid ){
int j;
for(j=0; j<nCol; j++){
if( sqlite3_stricmp(zLanguageid, aCol[j])==0 ){
int k;
for(k=j; k<nCol; k++) aCol[k] = aCol[k+1];
nCol--;
break;
}
}
}
}
}
if( rc!=SQLITE_OK ) goto fts3_init_out;
if( nCol==0 ){
assert( nString==0 );
aCol[0] = "content";
nString = 8;
nCol = 1;
}
if( pTokenizer==0 ){
rc = sqlite3Fts3InitTokenizer(pHash, "simple", &pTokenizer, pzErr);
if( rc!=SQLITE_OK ) goto fts3_init_out;
}
assert( pTokenizer );
rc = fts3PrefixParameter(zPrefix, &nIndex, &aIndex);
if( rc==SQLITE_ERROR ){
assert( zPrefix );
sqlite3Fts3ErrMsg(pzErr, "error parsing prefix parameter: %s", zPrefix);
}
if( rc!=SQLITE_OK ) goto fts3_init_out;
/* Allocate and populate the Fts3Table structure. */
nByte = sizeof(Fts3Table) + /* Fts3Table */
nCol * sizeof(char *) + /* azColumn */
nIndex * sizeof(struct Fts3Index) + /* aIndex */
nCol * sizeof(u8) + /* abNotindexed */
nName + /* zName */
nDb + /* zDb */
nString; /* Space for azColumn strings */
p = (Fts3Table*)sqlite3_malloc64(nByte);
if( p==0 ){
rc = SQLITE_NOMEM;
goto fts3_init_out;
}
memset(p, 0, nByte);
p->db = db;
p->nColumn = nCol;
p->nPendingData = 0;
p->azColumn = (char **)&p[1];
p->pTokenizer = pTokenizer;
p->nMaxPendingData = FTS3_MAX_PENDING_DATA;
p->bHasDocsize = (isFts4 && bNoDocsize==0);
p->bHasStat = (u8)isFts4;
p->bFts4 = (u8)isFts4;
p->bDescIdx = (u8)bDescIdx;
p->nAutoincrmerge = 0xff; /* 0xff means setting unknown */
p->zContentTbl = zContent;
p->zLanguageid = zLanguageid;
zContent = 0;
zLanguageid = 0;
TESTONLY( p->inTransaction = -1 );
TESTONLY( p->mxSavepoint = -1 );
p->aIndex = (struct Fts3Index *)&p->azColumn[nCol];
memcpy(p->aIndex, aIndex, sizeof(struct Fts3Index) * nIndex);
p->nIndex = nIndex;
for(i=0; i<nIndex; i++){
fts3HashInit(&p->aIndex[i].hPending, FTS3_HASH_STRING, 1);
}
p->abNotindexed = (u8 *)&p->aIndex[nIndex];
/* Fill in the zName and zDb fields of the vtab structure. */
zCsr = (char *)&p->abNotindexed[nCol];
p->zName = zCsr;
memcpy(zCsr, argv[2], nName);
zCsr += nName;
p->zDb = zCsr;
memcpy(zCsr, argv[1], nDb);
zCsr += nDb;
/* Fill in the azColumn array */
for(iCol=0; iCol<nCol; iCol++){
char *z;
int n = 0;
z = (char *)sqlite3Fts3NextToken(aCol[iCol], &n);
if( n>0 ){
memcpy(zCsr, z, n);
}
zCsr[n] = '\0';
sqlite3Fts3Dequote(zCsr);
p->azColumn[iCol] = zCsr;
zCsr += n+1;
assert( zCsr <= &((char *)p)[nByte] );
}
/* Fill in the abNotindexed array */
for(iCol=0; iCol<nCol; iCol++){
int n = (int)strlen(p->azColumn[iCol]);
for(i=0; i<nNotindexed; i++){
char *zNot = azNotindexed[i];
if( zNot && n==(int)strlen(zNot)
&& 0==sqlite3_strnicmp(p->azColumn[iCol], zNot, n)
){
p->abNotindexed[iCol] = 1;
sqlite3_free(zNot);
azNotindexed[i] = 0;
}
}
}
for(i=0; i<nNotindexed; i++){
if( azNotindexed[i] ){
sqlite3Fts3ErrMsg(pzErr, "no such column: %s", azNotindexed[i]);
rc = SQLITE_ERROR;
}
}
if( rc==SQLITE_OK && (zCompress==0)!=(zUncompress==0) ){
char const *zMiss = (zCompress==0 ? "compress" : "uncompress");
rc = SQLITE_ERROR;
sqlite3Fts3ErrMsg(pzErr, "missing %s parameter in fts4 constructor", zMiss);
}
p->zReadExprlist = fts3ReadExprList(p, zUncompress, &rc);
p->zWriteExprlist = fts3WriteExprList(p, zCompress, &rc);
if( rc!=SQLITE_OK ) goto fts3_init_out;
/* If this is an xCreate call, create the underlying tables in the
** database. TODO: For xConnect(), it could verify that said tables exist.
*/
if( isCreate ){
rc = fts3CreateTables(p);
}
/* Check to see if a legacy fts3 table has been "upgraded" by the
** addition of a %_stat table so that it can use incremental merge.
*/
if( !isFts4 && !isCreate ){
p->bHasStat = 2;
}
/* Figure out the page-size for the database. This is required in order to
** estimate the cost of loading large doclists from the database. */
fts3DatabasePageSize(&rc, p);
p->nNodeSize = p->nPgsz-35;
#if defined(SQLITE_DEBUG)||defined(SQLITE_TEST)
p->nMergeCount = FTS3_MERGE_COUNT;
#endif
/* Declare the table schema to SQLite. */
fts3DeclareVtab(&rc, p);
fts3_init_out:
sqlite3_free(zPrefix);
sqlite3_free(aIndex);
sqlite3_free(zCompress);
sqlite3_free(zUncompress);
sqlite3_free(zContent);
sqlite3_free(zLanguageid);
for(i=0; i<nNotindexed; i++) sqlite3_free(azNotindexed[i]);
sqlite3_free((void *)aCol);
sqlite3_free((void *)azNotindexed);
if( rc!=SQLITE_OK ){
if( p ){
fts3DisconnectMethod((sqlite3_vtab *)p);
}else if( pTokenizer ){
pTokenizer->pModule->xDestroy(pTokenizer);
}
}else{
assert( p->pSegments==0 );
*ppVTab = &p->base;
}
return rc;
}
/*
** The xConnect() and xCreate() methods for the virtual table. All the
** work is done in function fts3InitVtab().
*/
static int fts3ConnectMethod(
sqlite3 *db, /* Database connection */
void *pAux, /* Pointer to tokenizer hash table */
int argc, /* Number of elements in argv array */
const char * const *argv, /* xCreate/xConnect argument array */
sqlite3_vtab **ppVtab, /* OUT: New sqlite3_vtab object */
char **pzErr /* OUT: sqlite3_malloc'd error message */
){
return fts3InitVtab(0, db, pAux, argc, argv, ppVtab, pzErr);
}
static int fts3CreateMethod(
sqlite3 *db, /* Database connection */
void *pAux, /* Pointer to tokenizer hash table */
int argc, /* Number of elements in argv array */
const char * const *argv, /* xCreate/xConnect argument array */
sqlite3_vtab **ppVtab, /* OUT: New sqlite3_vtab object */
char **pzErr /* OUT: sqlite3_malloc'd error message */
){
return fts3InitVtab(1, db, pAux, argc, argv, ppVtab, pzErr);
}
/*
** Set the pIdxInfo->estimatedRows variable to nRow. Unless this
** extension is currently being used by a version of SQLite too old to
** support estimatedRows. In that case this function is a no-op.
*/
static void fts3SetEstimatedRows(sqlite3_index_info *pIdxInfo, i64 nRow){
#if SQLITE_VERSION_NUMBER>=3008002
if( sqlite3_libversion_number()>=3008002 ){
pIdxInfo->estimatedRows = nRow;
}
#endif
}
/*
** Set the SQLITE_INDEX_SCAN_UNIQUE flag in pIdxInfo->flags. Unless this
** extension is currently being used by a version of SQLite too old to
** support index-info flags. In that case this function is a no-op.
*/
static void fts3SetUniqueFlag(sqlite3_index_info *pIdxInfo){
#if SQLITE_VERSION_NUMBER>=3008012
if( sqlite3_libversion_number()>=3008012 ){
pIdxInfo->idxFlags |= SQLITE_INDEX_SCAN_UNIQUE;
}
#endif
}
/*
** Implementation of the xBestIndex method for FTS3 tables. There
** are three possible strategies, in order of preference:
**
** 1. Direct lookup by rowid or docid.
** 2. Full-text search using a MATCH operator on a non-docid column.
** 3. Linear scan of %_content table.
*/
static int fts3BestIndexMethod(sqlite3_vtab *pVTab, sqlite3_index_info *pInfo){
Fts3Table *p = (Fts3Table *)pVTab;
int i; /* Iterator variable */
int iCons = -1; /* Index of constraint to use */
int iLangidCons = -1; /* Index of langid=x constraint, if present */
int iDocidGe = -1; /* Index of docid>=x constraint, if present */
int iDocidLe = -1; /* Index of docid<=x constraint, if present */
int iIdx;
if( p->bLock ){
return SQLITE_ERROR;
}
/* By default use a full table scan. This is an expensive option,
** so search through the constraints to see if a more efficient
** strategy is possible.
*/
pInfo->idxNum = FTS3_FULLSCAN_SEARCH;
pInfo->estimatedCost = 5000000;
for(i=0; i<pInfo->nConstraint; i++){
int bDocid; /* True if this constraint is on docid */
struct sqlite3_index_constraint *pCons = &pInfo->aConstraint[i];
if( pCons->usable==0 ){
if( pCons->op==SQLITE_INDEX_CONSTRAINT_MATCH ){
/* There exists an unusable MATCH constraint. This means that if
** the planner does elect to use the results of this call as part
** of the overall query plan the user will see an "unable to use
** function MATCH in the requested context" error. To discourage
** this, return a very high cost here. */
pInfo->idxNum = FTS3_FULLSCAN_SEARCH;
pInfo->estimatedCost = 1e50;
fts3SetEstimatedRows(pInfo, ((sqlite3_int64)1) << 50);
return SQLITE_OK;
}
continue;
}
bDocid = (pCons->iColumn<0 || pCons->iColumn==p->nColumn+1);
/* A direct lookup on the rowid or docid column. Assign a cost of 1.0. */
if( iCons<0 && pCons->op==SQLITE_INDEX_CONSTRAINT_EQ && bDocid ){
pInfo->idxNum = FTS3_DOCID_SEARCH;
pInfo->estimatedCost = 1.0;
iCons = i;
}
/* A MATCH constraint. Use a full-text search.
**
** If there is more than one MATCH constraint available, use the first
** one encountered. If there is both a MATCH constraint and a direct
** rowid/docid lookup, prefer the MATCH strategy. This is done even
** though the rowid/docid lookup is faster than a MATCH query, selecting
** it would lead to an "unable to use function MATCH in the requested
** context" error.
*/
if( pCons->op==SQLITE_INDEX_CONSTRAINT_MATCH
&& pCons->iColumn>=0 && pCons->iColumn<=p->nColumn
){
pInfo->idxNum = FTS3_FULLTEXT_SEARCH + pCons->iColumn;
pInfo->estimatedCost = 2.0;
iCons = i;
}
/* Equality constraint on the langid column */
if( pCons->op==SQLITE_INDEX_CONSTRAINT_EQ
&& pCons->iColumn==p->nColumn + 2
){
iLangidCons = i;
}
if( bDocid ){
switch( pCons->op ){
case SQLITE_INDEX_CONSTRAINT_GE:
case SQLITE_INDEX_CONSTRAINT_GT:
iDocidGe = i;
break;
case SQLITE_INDEX_CONSTRAINT_LE:
case SQLITE_INDEX_CONSTRAINT_LT:
iDocidLe = i;
break;
}
}
}
/* If using a docid=? or rowid=? strategy, set the UNIQUE flag. */
if( pInfo->idxNum==FTS3_DOCID_SEARCH ) fts3SetUniqueFlag(pInfo);
iIdx = 1;
if( iCons>=0 ){
pInfo->aConstraintUsage[iCons].argvIndex = iIdx++;
pInfo->aConstraintUsage[iCons].omit = 1;
}
if( iLangidCons>=0 ){
pInfo->idxNum |= FTS3_HAVE_LANGID;
pInfo->aConstraintUsage[iLangidCons].argvIndex = iIdx++;
}
if( iDocidGe>=0 ){
pInfo->idxNum |= FTS3_HAVE_DOCID_GE;
pInfo->aConstraintUsage[iDocidGe].argvIndex = iIdx++;
}
if( iDocidLe>=0 ){
pInfo->idxNum |= FTS3_HAVE_DOCID_LE;
pInfo->aConstraintUsage[iDocidLe].argvIndex = iIdx++;
}
/* Regardless of the strategy selected, FTS can deliver rows in rowid (or
** docid) order. Both ascending and descending are possible.
*/
if( pInfo->nOrderBy==1 ){
struct sqlite3_index_orderby *pOrder = &pInfo->aOrderBy[0];
if( pOrder->iColumn<0 || pOrder->iColumn==p->nColumn+1 ){
if( pOrder->desc ){
pInfo->idxStr = "DESC";
}else{
pInfo->idxStr = "ASC";
}
pInfo->orderByConsumed = 1;
}
}
assert( p->pSegments==0 );
return SQLITE_OK;
}
/*
** Implementation of xOpen method.
*/
static int fts3OpenMethod(sqlite3_vtab *pVTab, sqlite3_vtab_cursor **ppCsr){
sqlite3_vtab_cursor *pCsr; /* Allocated cursor */
UNUSED_PARAMETER(pVTab);
/* Allocate a buffer large enough for an Fts3Cursor structure. If the
** allocation succeeds, zero it and return SQLITE_OK. Otherwise,
** if the allocation fails, return SQLITE_NOMEM.
*/
*ppCsr = pCsr = (sqlite3_vtab_cursor *)sqlite3_malloc(sizeof(Fts3Cursor));
if( !pCsr ){
return SQLITE_NOMEM;
}
memset(pCsr, 0, sizeof(Fts3Cursor));
return SQLITE_OK;
}
/*
** Finalize the statement handle at pCsr->pStmt.
**
** Or, if that statement handle is one created by fts3CursorSeekStmt(),
** and the Fts3Table.pSeekStmt slot is currently NULL, save the statement
** pointer there instead of finalizing it.
*/
static void fts3CursorFinalizeStmt(Fts3Cursor *pCsr){
if( pCsr->bSeekStmt ){
Fts3Table *p = (Fts3Table *)pCsr->base.pVtab;
if( p->pSeekStmt==0 ){
p->pSeekStmt = pCsr->pStmt;
sqlite3_reset(pCsr->pStmt);
pCsr->pStmt = 0;
}
pCsr->bSeekStmt = 0;
}
sqlite3_finalize(pCsr->pStmt);
}
/*
** Free all resources currently held by the cursor passed as the only
** argument.
*/
static void fts3ClearCursor(Fts3Cursor *pCsr){
fts3CursorFinalizeStmt(pCsr);
sqlite3Fts3FreeDeferredTokens(pCsr);
sqlite3_free(pCsr->aDoclist);
sqlite3Fts3MIBufferFree(pCsr->pMIBuffer);
sqlite3Fts3ExprFree(pCsr->pExpr);
memset(&(&pCsr->base)[1], 0, sizeof(Fts3Cursor)-sizeof(sqlite3_vtab_cursor));
}
/*
** Close the cursor. For additional information see the documentation
** on the xClose method of the virtual table interface.
*/
static int fts3CloseMethod(sqlite3_vtab_cursor *pCursor){
Fts3Cursor *pCsr = (Fts3Cursor *)pCursor;
assert( ((Fts3Table *)pCsr->base.pVtab)->pSegments==0 );
fts3ClearCursor(pCsr);
assert( ((Fts3Table *)pCsr->base.pVtab)->pSegments==0 );
sqlite3_free(pCsr);
return SQLITE_OK;
}
/*
** If pCsr->pStmt has not been prepared (i.e. if pCsr->pStmt==0), then
** compose and prepare an SQL statement of the form:
**
** "SELECT <columns> FROM %_content WHERE rowid = ?"
**
** (or the equivalent for a content=xxx table) and set pCsr->pStmt to
** it. If an error occurs, return an SQLite error code.
*/
static int fts3CursorSeekStmt(Fts3Cursor *pCsr){
int rc = SQLITE_OK;
if( pCsr->pStmt==0 ){
Fts3Table *p = (Fts3Table *)pCsr->base.pVtab;
char *zSql;
if( p->pSeekStmt ){
pCsr->pStmt = p->pSeekStmt;
p->pSeekStmt = 0;
}else{
zSql = sqlite3_mprintf("SELECT %s WHERE rowid = ?", p->zReadExprlist);
if( !zSql ) return SQLITE_NOMEM;
p->bLock++;
rc = sqlite3_prepare_v3(
p->db, zSql,-1,SQLITE_PREPARE_PERSISTENT,&pCsr->pStmt,0
);
p->bLock--;
sqlite3_free(zSql);
}
if( rc==SQLITE_OK ) pCsr->bSeekStmt = 1;
}
return rc;
}
/*
** Position the pCsr->pStmt statement so that it is on the row
** of the %_content table that contains the last match. Return
** SQLITE_OK on success.
*/
static int fts3CursorSeek(sqlite3_context *pContext, Fts3Cursor *pCsr){
int rc = SQLITE_OK;
if( pCsr->isRequireSeek ){
rc = fts3CursorSeekStmt(pCsr);
if( rc==SQLITE_OK ){
Fts3Table *pTab = (Fts3Table*)pCsr->base.pVtab;
pTab->bLock++;
sqlite3_bind_int64(pCsr->pStmt, 1, pCsr->iPrevId);
pCsr->isRequireSeek = 0;
if( SQLITE_ROW==sqlite3_step(pCsr->pStmt) ){
pTab->bLock--;
return SQLITE_OK;
}else{
pTab->bLock--;
rc = sqlite3_reset(pCsr->pStmt);
if( rc==SQLITE_OK && ((Fts3Table *)pCsr->base.pVtab)->zContentTbl==0 ){
/* If no row was found and no error has occurred, then the %_content
** table is missing a row that is present in the full-text index.
** The data structures are corrupt. */
rc = FTS_CORRUPT_VTAB;
pCsr->isEof = 1;
}
}
}
}
if( rc!=SQLITE_OK && pContext ){
sqlite3_result_error_code(pContext, rc);
}
return rc;
}
/*
** This function is used to process a single interior node when searching
** a b-tree for a term or term prefix. The node data is passed to this
** function via the zNode/nNode parameters. The term to search for is
** passed in zTerm/nTerm.
**
** If piFirst is not NULL, then this function sets *piFirst to the blockid
** of the child node that heads the sub-tree that may contain the term.
**
** If piLast is not NULL, then *piLast is set to the right-most child node
** that heads a sub-tree that may contain a term for which zTerm/nTerm is
** a prefix.
**
** If an OOM error occurs, SQLITE_NOMEM is returned. Otherwise, SQLITE_OK.
*/
static int fts3ScanInteriorNode(
const char *zTerm, /* Term to select leaves for */
int nTerm, /* Size of term zTerm in bytes */
const char *zNode, /* Buffer containing segment interior node */
int nNode, /* Size of buffer at zNode */
sqlite3_int64 *piFirst, /* OUT: Selected child node */
sqlite3_int64 *piLast /* OUT: Selected child node */
){
int rc = SQLITE_OK; /* Return code */
const char *zCsr = zNode; /* Cursor to iterate through node */
const char *zEnd = &zCsr[nNode];/* End of interior node buffer */
char *zBuffer = 0; /* Buffer to load terms into */
i64 nAlloc = 0; /* Size of allocated buffer */
int isFirstTerm = 1; /* True when processing first term on page */
u64 iChild; /* Block id of child node to descend to */
int nBuffer = 0; /* Total term size */
/* Skip over the 'height' varint that occurs at the start of every
** interior node. Then load the blockid of the left-child of the b-tree
** node into variable iChild.
**
** Even if the data structure on disk is corrupted, this (reading two
** varints from the buffer) does not risk an overread. If zNode is a
** root node, then the buffer comes from a SELECT statement. SQLite does
** not make this guarantee explicitly, but in practice there are always
** either more than 20 bytes of allocated space following the nNode bytes of
** contents, or two zero bytes. Or, if the node is read from the %_segments
** table, then there are always 20 bytes of zeroed padding following the
** nNode bytes of content (see sqlite3Fts3ReadBlock() for details).
*/
zCsr += sqlite3Fts3GetVarintU(zCsr, &iChild);
zCsr += sqlite3Fts3GetVarintU(zCsr, &iChild);
if( zCsr>zEnd ){
return FTS_CORRUPT_VTAB;
}
while( zCsr<zEnd && (piFirst || piLast) ){
int cmp; /* memcmp() result */
int nSuffix; /* Size of term suffix */
int nPrefix = 0; /* Size of term prefix */
/* Load the next term on the node into zBuffer. Use realloc() to expand
** the size of zBuffer if required. */
if( !isFirstTerm ){
zCsr += fts3GetVarint32(zCsr, &nPrefix);
if( nPrefix>nBuffer ){
rc = FTS_CORRUPT_VTAB;
goto finish_scan;
}
}
isFirstTerm = 0;
zCsr += fts3GetVarint32(zCsr, &nSuffix);
assert( nPrefix>=0 && nSuffix>=0 );
if( nPrefix>zCsr-zNode || nSuffix>zEnd-zCsr || nSuffix==0 ){
rc = FTS_CORRUPT_VTAB;
goto finish_scan;
}
if( (i64)nPrefix+nSuffix>nAlloc ){
char *zNew;
nAlloc = ((i64)nPrefix+nSuffix) * 2;
zNew = (char *)sqlite3_realloc64(zBuffer, nAlloc);
if( !zNew ){
rc = SQLITE_NOMEM;
goto finish_scan;
}
zBuffer = zNew;
}
assert( zBuffer );
memcpy(&zBuffer[nPrefix], zCsr, nSuffix);
nBuffer = nPrefix + nSuffix;
zCsr += nSuffix;
/* Compare the term we are searching for with the term just loaded from
** the interior node. If the specified term is greater than or equal
** to the term from the interior node, then all terms on the sub-tree
** headed by node iChild are smaller than zTerm. No need to search
** iChild.
**
** If the interior node term is larger than the specified term, then
** the tree headed by iChild may contain the specified term.
*/
cmp = memcmp(zTerm, zBuffer, (nBuffer>nTerm ? nTerm : nBuffer));
if( piFirst && (cmp<0 || (cmp==0 && nBuffer>nTerm)) ){
*piFirst = (i64)iChild;
piFirst = 0;
}
if( piLast && cmp<0 ){
*piLast = (i64)iChild;
piLast = 0;
}
iChild++;
};
if( piFirst ) *piFirst = (i64)iChild;
if( piLast ) *piLast = (i64)iChild;
finish_scan:
sqlite3_free(zBuffer);
return rc;
}
/*
** The buffer pointed to by argument zNode (size nNode bytes) contains an
** interior node of a b-tree segment. The zTerm buffer (size nTerm bytes)
** contains a term. This function searches the sub-tree headed by the zNode
** node for the range of leaf nodes that may contain the specified term
** or terms for which the specified term is a prefix.
**
** If piLeaf is not NULL, then *piLeaf is set to the blockid of the
** left-most leaf node in the tree that may contain the specified term.
** If piLeaf2 is not NULL, then *piLeaf2 is set to the blockid of the
** right-most leaf node that may contain a term for which the specified
** term is a prefix.
**
** It is possible that the range of returned leaf nodes does not contain
** the specified term or any terms for which it is a prefix. However, if the
** segment does contain any such terms, they are stored within the identified
** range. Because this function only inspects interior segment nodes (and
** never loads leaf nodes into memory), it is not possible to be sure.
**
** If an error occurs, an error code other than SQLITE_OK is returned.
*/
static int fts3SelectLeaf(
Fts3Table *p, /* Virtual table handle */
const char *zTerm, /* Term to select leaves for */
int nTerm, /* Size of term zTerm in bytes */
const char *zNode, /* Buffer containing segment interior node */
int nNode, /* Size of buffer at zNode */
sqlite3_int64 *piLeaf, /* Selected leaf node */
sqlite3_int64 *piLeaf2 /* Selected leaf node */
){
int rc = SQLITE_OK; /* Return code */
int iHeight; /* Height of this node in tree */
assert( piLeaf || piLeaf2 );
fts3GetVarint32(zNode, &iHeight);
rc = fts3ScanInteriorNode(zTerm, nTerm, zNode, nNode, piLeaf, piLeaf2);
assert_fts3_nc( !piLeaf2 || !piLeaf || rc!=SQLITE_OK || (*piLeaf<=*piLeaf2) );
if( rc==SQLITE_OK && iHeight>1 ){
char *zBlob = 0; /* Blob read from %_segments table */
int nBlob = 0; /* Size of zBlob in bytes */
if( piLeaf && piLeaf2 && (*piLeaf!=*piLeaf2) ){
rc = sqlite3Fts3ReadBlock(p, *piLeaf, &zBlob, &nBlob, 0);
if( rc==SQLITE_OK ){
rc = fts3SelectLeaf(p, zTerm, nTerm, zBlob, nBlob, piLeaf, 0);
}
sqlite3_free(zBlob);
piLeaf = 0;
zBlob = 0;
}
if( rc==SQLITE_OK ){
rc = sqlite3Fts3ReadBlock(p, piLeaf?*piLeaf:*piLeaf2, &zBlob, &nBlob, 0);
}
if( rc==SQLITE_OK ){
int iNewHeight = 0;
fts3GetVarint32(zBlob, &iNewHeight);
if( iNewHeight>=iHeight ){
rc = FTS_CORRUPT_VTAB;
}else{
rc = fts3SelectLeaf(p, zTerm, nTerm, zBlob, nBlob, piLeaf, piLeaf2);
}
}
sqlite3_free(zBlob);
}
return rc;
}
/*
** This function is used to create delta-encoded serialized lists of FTS3
** varints. Each call to this function appends a single varint to a list.
*/
static void fts3PutDeltaVarint(
char **pp, /* IN/OUT: Output pointer */
sqlite3_int64 *piPrev, /* IN/OUT: Previous value written to list */
sqlite3_int64 iVal /* Write this value to the list */
){
assert_fts3_nc( iVal-*piPrev > 0 || (*piPrev==0 && iVal==0) );
*pp += sqlite3Fts3PutVarint(*pp, iVal-*piPrev);
*piPrev = iVal;
}
/*
** When this function is called, *ppPoslist is assumed to point to the
** start of a position-list. After it returns, *ppPoslist points to the
** first byte after the position-list.
**
** A position list is list of positions (delta encoded) and columns for
** a single document record of a doclist. So, in other words, this
** routine advances *ppPoslist so that it points to the next docid in
** the doclist, or to the first byte past the end of the doclist.
**
** If pp is not NULL, then the contents of the position list are copied
** to *pp. *pp is set to point to the first byte past the last byte copied
** before this function returns.
*/
static void fts3PoslistCopy(char **pp, char **ppPoslist){
char *pEnd = *ppPoslist;
char c = 0;
/* The end of a position list is marked by a zero encoded as an FTS3
** varint. A single POS_END (0) byte. Except, if the 0 byte is preceded by
** a byte with the 0x80 bit set, then it is not a varint 0, but the tail
** of some other, multi-byte, value.
**
** The following while-loop moves pEnd to point to the first byte that is not
** immediately preceded by a byte with the 0x80 bit set. Then increments
** pEnd once more so that it points to the byte immediately following the
** last byte in the position-list.
*/
while( *pEnd | c ){
c = *pEnd++ & 0x80;
testcase( c!=0 && (*pEnd)==0 );
}
pEnd++; /* Advance past the POS_END terminator byte */
if( pp ){
int n = (int)(pEnd - *ppPoslist);
char *p = *pp;
memcpy(p, *ppPoslist, n);
p += n;
*pp = p;
}
*ppPoslist = pEnd;
}
/*
** When this function is called, *ppPoslist is assumed to point to the
** start of a column-list. After it returns, *ppPoslist points to the
** to the terminator (POS_COLUMN or POS_END) byte of the column-list.
**
** A column-list is list of delta-encoded positions for a single column
** within a single document within a doclist.
**
** The column-list is terminated either by a POS_COLUMN varint (1) or
** a POS_END varint (0). This routine leaves *ppPoslist pointing to
** the POS_COLUMN or POS_END that terminates the column-list.
**
** If pp is not NULL, then the contents of the column-list are copied
** to *pp. *pp is set to point to the first byte past the last byte copied
** before this function returns. The POS_COLUMN or POS_END terminator
** is not copied into *pp.
*/
static void fts3ColumnlistCopy(char **pp, char **ppPoslist){
char *pEnd = *ppPoslist;
char c = 0;
/* A column-list is terminated by either a 0x01 or 0x00 byte that is
** not part of a multi-byte varint.
*/
while( 0xFE & (*pEnd | c) ){
c = *pEnd++ & 0x80;
testcase( c!=0 && ((*pEnd)&0xfe)==0 );
}
if( pp ){
int n = (int)(pEnd - *ppPoslist);
char *p = *pp;
memcpy(p, *ppPoslist, n);
p += n;
*pp = p;
}
*ppPoslist = pEnd;
}
/*
** Value used to signify the end of an position-list. This must be
** as large or larger than any value that might appear on the
** position-list, even a position list that has been corrupted.
*/
#define POSITION_LIST_END LARGEST_INT64
/*
** This function is used to help parse position-lists. When this function is
** called, *pp may point to the start of the next varint in the position-list
** being parsed, or it may point to 1 byte past the end of the position-list
** (in which case **pp will be a terminator bytes POS_END (0) or
** (1)).
**
** If *pp points past the end of the current position-list, set *pi to
** POSITION_LIST_END and return. Otherwise, read the next varint from *pp,
** increment the current value of *pi by the value read, and set *pp to
** point to the next value before returning.
**
** Before calling this routine *pi must be initialized to the value of
** the previous position, or zero if we are reading the first position
** in the position-list. Because positions are delta-encoded, the value
** of the previous position is needed in order to compute the value of
** the next position.
*/
static void fts3ReadNextPos(
char **pp, /* IN/OUT: Pointer into position-list buffer */
sqlite3_int64 *pi /* IN/OUT: Value read from position-list */
){
if( (**pp)&0xFE ){
int iVal;
*pp += fts3GetVarint32((*pp), &iVal);
*pi += iVal;
*pi -= 2;
}else{
*pi = POSITION_LIST_END;
}
}
/*
** If parameter iCol is not 0, write an POS_COLUMN (1) byte followed by
** the value of iCol encoded as a varint to *pp. This will start a new
** column list.
**
** Set *pp to point to the byte just after the last byte written before
** returning (do not modify it if iCol==0). Return the total number of bytes
** written (0 if iCol==0).
*/
static int fts3PutColNumber(char **pp, int iCol){
int n = 0; /* Number of bytes written */
if( iCol ){
char *p = *pp; /* Output pointer */
n = 1 + sqlite3Fts3PutVarint(&p[1], iCol);
*p = 0x01;
*pp = &p[n];
}
return n;
}
/*
** Compute the union of two position lists. The output written
** into *pp contains all positions of both *pp1 and *pp2 in sorted
** order and with any duplicates removed. All pointers are
** updated appropriately. The caller is responsible for insuring
** that there is enough space in *pp to hold the complete output.
*/
static int fts3PoslistMerge(
char **pp, /* Output buffer */
char **pp1, /* Left input list */
char **pp2 /* Right input list */
){
char *p = *pp;
char *p1 = *pp1;
char *p2 = *pp2;
while( *p1 || *p2 ){
int iCol1; /* The current column index in pp1 */
int iCol2; /* The current column index in pp2 */
if( *p1==POS_COLUMN ){
fts3GetVarint32(&p1[1], &iCol1);
if( iCol1==0 ) return FTS_CORRUPT_VTAB;
}
else if( *p1==POS_END ) iCol1 = 0x7fffffff;
else iCol1 = 0;
if( *p2==POS_COLUMN ){
fts3GetVarint32(&p2[1], &iCol2);
if( iCol2==0 ) return FTS_CORRUPT_VTAB;
}
else if( *p2==POS_END ) iCol2 = 0x7fffffff;
else iCol2 = 0;
if( iCol1==iCol2 ){
sqlite3_int64 i1 = 0; /* Last position from pp1 */
sqlite3_int64 i2 = 0; /* Last position from pp2 */
sqlite3_int64 iPrev = 0;
int n = fts3PutColNumber(&p, iCol1);
p1 += n;
p2 += n;
/* At this point, both p1 and p2 point to the start of column-lists
** for the same column (the column with index iCol1 and iCol2).
** A column-list is a list of non-negative delta-encoded varints, each
** incremented by 2 before being stored. Each list is terminated by a
** POS_END (0) or POS_COLUMN (1). The following block merges the two lists
** and writes the results to buffer p. p is left pointing to the byte
** after the list written. No terminator (POS_END or POS_COLUMN) is
** written to the output.
*/
fts3GetDeltaVarint(&p1, &i1);
fts3GetDeltaVarint(&p2, &i2);
if( i1<2 || i2<2 ){
break;
}
do {
fts3PutDeltaVarint(&p, &iPrev, (i1<i2) ? i1 : i2);
iPrev -= 2;
if( i1==i2 ){
fts3ReadNextPos(&p1, &i1);
fts3ReadNextPos(&p2, &i2);
}else if( i1<i2 ){
fts3ReadNextPos(&p1, &i1);
}else{
fts3ReadNextPos(&p2, &i2);
}
}while( i1!=POSITION_LIST_END || i2!=POSITION_LIST_END );
}else if( iCol1<iCol2 ){
p1 += fts3PutColNumber(&p, iCol1);
fts3ColumnlistCopy(&p, &p1);
}else{
p2 += fts3PutColNumber(&p, iCol2);
fts3ColumnlistCopy(&p, &p2);
}
}
*p++ = POS_END;
*pp = p;
*pp1 = p1 + 1;
*pp2 = p2 + 1;
return SQLITE_OK;
}
/*
** This function is used to merge two position lists into one. When it is
** called, *pp1 and *pp2 must both point to position lists. A position-list is
** the part of a doclist that follows each document id. For example, if a row
** contains:
**
** 'a b c'|'x y z'|'a b b a'
**
** Then the position list for this row for token 'b' would consist of:
**
** 0x02 0x01 0x02 0x03 0x03 0x00
**
** When this function returns, both *pp1 and *pp2 are left pointing to the
** byte following the 0x00 terminator of their respective position lists.
**
** If isSaveLeft is 0, an entry is added to the output position list for
** each position in *pp2 for which there exists one or more positions in
** *pp1 so that (pos(*pp2)>pos(*pp1) && pos(*pp2)-pos(*pp1)<=nToken). i.e.
** when the *pp1 token appears before the *pp2 token, but not more than nToken
** slots before it.
**
** e.g. nToken==1 searches for adjacent positions.
*/
static int fts3PoslistPhraseMerge(
char **pp, /* IN/OUT: Preallocated output buffer */
int nToken, /* Maximum difference in token positions */
int isSaveLeft, /* Save the left position */
int isExact, /* If *pp1 is exactly nTokens before *pp2 */
char **pp1, /* IN/OUT: Left input list */
char **pp2 /* IN/OUT: Right input list */
){
char *p = *pp;
char *p1 = *pp1;
char *p2 = *pp2;
int iCol1 = 0;
int iCol2 = 0;
/* Never set both isSaveLeft and isExact for the same invocation. */
assert( isSaveLeft==0 || isExact==0 );
assert_fts3_nc( p!=0 && *p1!=0 && *p2!=0 );
if( *p1==POS_COLUMN ){
p1++;
p1 += fts3GetVarint32(p1, &iCol1);
}
if( *p2==POS_COLUMN ){
p2++;
p2 += fts3GetVarint32(p2, &iCol2);
}
while( 1 ){
if( iCol1==iCol2 ){
char *pSave = p;
sqlite3_int64 iPrev = 0;
sqlite3_int64 iPos1 = 0;
sqlite3_int64 iPos2 = 0;
if( iCol1 ){
*p++ = POS_COLUMN;
p += sqlite3Fts3PutVarint(p, iCol1);
}
fts3GetDeltaVarint(&p1, &iPos1); iPos1 -= 2;
fts3GetDeltaVarint(&p2, &iPos2); iPos2 -= 2;
if( iPos1<0 || iPos2<0 ) break;
while( 1 ){
if( iPos2==iPos1+nToken
|| (isExact==0 && iPos2>iPos1 && iPos2<=iPos1+nToken)
){
sqlite3_int64 iSave;
iSave = isSaveLeft ? iPos1 : iPos2;
fts3PutDeltaVarint(&p, &iPrev, iSave+2); iPrev -= 2;
pSave = 0;
assert( p );
}
if( (!isSaveLeft && iPos2<=(iPos1+nToken)) || iPos2<=iPos1 ){
if( (*p2&0xFE)==0 ) break;
fts3GetDeltaVarint(&p2, &iPos2); iPos2 -= 2;
}else{
if( (*p1&0xFE)==0 ) break;
fts3GetDeltaVarint(&p1, &iPos1); iPos1 -= 2;
}
}
if( pSave ){
assert( pp && p );
p = pSave;
}
fts3ColumnlistCopy(0, &p1);
fts3ColumnlistCopy(0, &p2);
assert( (*p1&0xFE)==0 && (*p2&0xFE)==0 );
if( 0==*p1 || 0==*p2 ) break;
p1++;
p1 += fts3GetVarint32(p1, &iCol1);
p2++;
p2 += fts3GetVarint32(p2, &iCol2);
}
/* Advance pointer p1 or p2 (whichever corresponds to the smaller of
** iCol1 and iCol2) so that it points to either the 0x00 that marks the
** end of the position list, or the 0x01 that precedes the next
** column-number in the position list.
*/
else if( iCol1<iCol2 ){
fts3ColumnlistCopy(0, &p1);
if( 0==*p1 ) break;
p1++;
p1 += fts3GetVarint32(p1, &iCol1);
}else{
fts3ColumnlistCopy(0, &p2);
if( 0==*p2 ) break;
p2++;
p2 += fts3GetVarint32(p2, &iCol2);
}
}
fts3PoslistCopy(0, &p2);
fts3PoslistCopy(0, &p1);
*pp1 = p1;
*pp2 = p2;
if( *pp==p ){
return 0;
}
*p++ = 0x00;
*pp = p;
return 1;
}
/*
** Merge two position-lists as required by the NEAR operator. The argument
** position lists correspond to the left and right phrases of an expression
** like:
**
** "phrase 1" NEAR "phrase number 2"
**
** Position list *pp1 corresponds to the left-hand side of the NEAR
** expression and *pp2 to the right. As usual, the indexes in the position
** lists are the offsets of the last token in each phrase (tokens "1" and "2"
** in the example above).
**
** The output position list - written to *pp - is a copy of *pp2 with those
** entries that are not sufficiently NEAR entries in *pp1 removed.
*/
static int fts3PoslistNearMerge(
char **pp, /* Output buffer */
char *aTmp, /* Temporary buffer space */
int nRight, /* Maximum difference in token positions */
int nLeft, /* Maximum difference in token positions */
char **pp1, /* IN/OUT: Left input list */
char **pp2 /* IN/OUT: Right input list */
){
char *p1 = *pp1;
char *p2 = *pp2;
char *pTmp1 = aTmp;
char *pTmp2;
char *aTmp2;
int res = 1;
fts3PoslistPhraseMerge(&pTmp1, nRight, 0, 0, pp1, pp2);
aTmp2 = pTmp2 = pTmp1;
*pp1 = p1;
*pp2 = p2;
fts3PoslistPhraseMerge(&pTmp2, nLeft, 1, 0, pp2, pp1);
if( pTmp1!=aTmp && pTmp2!=aTmp2 ){
fts3PoslistMerge(pp, &aTmp, &aTmp2);
}else if( pTmp1!=aTmp ){
fts3PoslistCopy(pp, &aTmp);
}else if( pTmp2!=aTmp2 ){
fts3PoslistCopy(pp, &aTmp2);
}else{
res = 0;
}
return res;
}
/*
** An instance of this function is used to merge together the (potentially
** large number of) doclists for each term that matches a prefix query.
** See function fts3TermSelectMerge() for details.
*/
typedef struct TermSelect TermSelect;
struct TermSelect {
char *aaOutput[16]; /* Malloc'd output buffers */
int anOutput[16]; /* Size each output buffer in bytes */
};
/*
** This function is used to read a single varint from a buffer. Parameter
** pEnd points 1 byte past the end of the buffer. When this function is
** called, if *pp points to pEnd or greater, then the end of the buffer
** has been reached. In this case *pp is set to 0 and the function returns.
**
** If *pp does not point to or past pEnd, then a single varint is read
** from *pp. *pp is then set to point 1 byte past the end of the read varint.
**
** If bDescIdx is false, the value read is added to *pVal before returning.
** If it is true, the value read is subtracted from *pVal before this
** function returns.
*/
static void fts3GetDeltaVarint3(
char **pp, /* IN/OUT: Point to read varint from */
char *pEnd, /* End of buffer */
int bDescIdx, /* True if docids are descending */
sqlite3_int64 *pVal /* IN/OUT: Integer value */
){
if( *pp>=pEnd ){
*pp = 0;
}else{
u64 iVal;
*pp += sqlite3Fts3GetVarintU(*pp, &iVal);
if( bDescIdx ){
*pVal = (i64)((u64)*pVal - iVal);
}else{
*pVal = (i64)((u64)*pVal + iVal);
}
}
}
/*
** This function is used to write a single varint to a buffer. The varint
** is written to *pp. Before returning, *pp is set to point 1 byte past the
** end of the value written.
**
** If *pbFirst is zero when this function is called, the value written to
** the buffer is that of parameter iVal.
**
** If *pbFirst is non-zero when this function is called, then the value
** written is either (iVal-*piPrev) (if bDescIdx is zero) or (*piPrev-iVal)
** (if bDescIdx is non-zero).
**
** Before returning, this function always sets *pbFirst to 1 and *piPrev
** to the value of parameter iVal.
*/
static void fts3PutDeltaVarint3(
char **pp, /* IN/OUT: Output pointer */
int bDescIdx, /* True for descending docids */
sqlite3_int64 *piPrev, /* IN/OUT: Previous value written to list */
int *pbFirst, /* IN/OUT: True after first int written */
sqlite3_int64 iVal /* Write this value to the list */
){
sqlite3_uint64 iWrite;
if( bDescIdx==0 || *pbFirst==0 ){
assert_fts3_nc( *pbFirst==0 || iVal>=*piPrev );
iWrite = (u64)iVal - (u64)*piPrev;
}else{
assert_fts3_nc( *piPrev>=iVal );
iWrite = (u64)*piPrev - (u64)iVal;
}
assert( *pbFirst || *piPrev==0 );
assert_fts3_nc( *pbFirst==0 || iWrite>0 );
*pp += sqlite3Fts3PutVarint(*pp, iWrite);
*piPrev = iVal;
*pbFirst = 1;
}
/*
** This macro is used by various functions that merge doclists. The two
** arguments are 64-bit docid values. If the value of the stack variable
** bDescDoclist is 0 when this macro is invoked, then it returns (i1-i2).
** Otherwise, (i2-i1).
**
** Using this makes it easier to write code that can merge doclists that are
** sorted in either ascending or descending order.
*/
/* #define DOCID_CMP(i1, i2) ((bDescDoclist?-1:1) * (i64)((u64)i1-i2)) */
#define DOCID_CMP(i1, i2) ((bDescDoclist?-1:1) * (i1>i2?1:((i1==i2)?0:-1)))
/*
** This function does an "OR" merge of two doclists (output contains all
** positions contained in either argument doclist). If the docids in the
** input doclists are sorted in ascending order, parameter bDescDoclist
** should be false. If they are sorted in ascending order, it should be
** passed a non-zero value.
**
** If no error occurs, *paOut is set to point at an sqlite3_malloc'd buffer
** containing the output doclist and SQLITE_OK is returned. In this case
** *pnOut is set to the number of bytes in the output doclist.
**
** If an error occurs, an SQLite error code is returned. The output values
** are undefined in this case.
*/
static int fts3DoclistOrMerge(
int bDescDoclist, /* True if arguments are desc */
char *a1, int n1, /* First doclist */
char *a2, int n2, /* Second doclist */
char **paOut, int *pnOut /* OUT: Malloc'd doclist */
){
int rc = SQLITE_OK;
sqlite3_int64 i1 = 0;
sqlite3_int64 i2 = 0;
sqlite3_int64 iPrev = 0;
char *pEnd1 = &a1[n1];
char *pEnd2 = &a2[n2];
char *p1 = a1;
char *p2 = a2;
char *p;
char *aOut;
int bFirstOut = 0;
*paOut = 0;
*pnOut = 0;
/* Allocate space for the output. Both the input and output doclists
** are delta encoded. If they are in ascending order (bDescDoclist==0),
** then the first docid in each list is simply encoded as a varint. For
** each subsequent docid, the varint stored is the difference between the
** current and previous docid (a positive number - since the list is in
** ascending order).
**
** The first docid written to the output is therefore encoded using the
** same number of bytes as it is in whichever of the input lists it is
** read from. And each subsequent docid read from the same input list
** consumes either the same or less bytes as it did in the input (since
** the difference between it and the previous value in the output must
** be a positive value less than or equal to the delta value read from
** the input list). The same argument applies to all but the first docid
** read from the 'other' list. And to the contents of all position lists
** that will be copied and merged from the input to the output.
**
** However, if the first docid copied to the output is a negative number,
** then the encoding of the first docid from the 'other' input list may
** be larger in the output than it was in the input (since the delta value
** may be a larger positive integer than the actual docid).
**
** The space required to store the output is therefore the sum of the
** sizes of the two inputs, plus enough space for exactly one of the input
** docids to grow.
**
** A symetric argument may be made if the doclists are in descending
** order.
*/
aOut = sqlite3_malloc64((i64)n1+n2+FTS3_VARINT_MAX-1+FTS3_BUFFER_PADDING);
if( !aOut ) return SQLITE_NOMEM;
p = aOut;
fts3GetDeltaVarint3(&p1, pEnd1, 0, &i1);
fts3GetDeltaVarint3(&p2, pEnd2, 0, &i2);
while( p1 || p2 ){
sqlite3_int64 iDiff = DOCID_CMP(i1, i2);
if( p2 && p1 && iDiff==0 ){
fts3PutDeltaVarint3(&p, bDescDoclist, &iPrev, &bFirstOut, i1);
rc = fts3PoslistMerge(&p, &p1, &p2);
if( rc ) break;
fts3GetDeltaVarint3(&p1, pEnd1, bDescDoclist, &i1);
fts3GetDeltaVarint3(&p2, pEnd2, bDescDoclist, &i2);
}else if( !p2 || (p1 && iDiff<0) ){
fts3PutDeltaVarint3(&p, bDescDoclist, &iPrev, &bFirstOut, i1);
fts3PoslistCopy(&p, &p1);
fts3GetDeltaVarint3(&p1, pEnd1, bDescDoclist, &i1);
}else{
fts3PutDeltaVarint3(&p, bDescDoclist, &iPrev, &bFirstOut, i2);
fts3PoslistCopy(&p, &p2);
fts3GetDeltaVarint3(&p2, pEnd2, bDescDoclist, &i2);
}
assert( (p-aOut)<=((p1?(p1-a1):n1)+(p2?(p2-a2):n2)+FTS3_VARINT_MAX-1) );
}
if( rc!=SQLITE_OK ){
sqlite3_free(aOut);
p = aOut = 0;
}else{
assert( (p-aOut)<=n1+n2+FTS3_VARINT_MAX-1 );
memset(&aOut[(p-aOut)], 0, FTS3_BUFFER_PADDING);
}
*paOut = aOut;
*pnOut = (int)(p-aOut);
return rc;
}
/*
** This function does a "phrase" merge of two doclists. In a phrase merge,
** the output contains a copy of each position from the right-hand input
** doclist for which there is a position in the left-hand input doclist
** exactly nDist tokens before it.
**
** If the docids in the input doclists are sorted in ascending order,
** parameter bDescDoclist should be false. If they are sorted in ascending
** order, it should be passed a non-zero value.
**
** The right-hand input doclist is overwritten by this function.
*/
static int fts3DoclistPhraseMerge(
int bDescDoclist, /* True if arguments are desc */
int nDist, /* Distance from left to right (1=adjacent) */
char *aLeft, int nLeft, /* Left doclist */
char **paRight, int *pnRight /* IN/OUT: Right/output doclist */
){
sqlite3_int64 i1 = 0;
sqlite3_int64 i2 = 0;
sqlite3_int64 iPrev = 0;
char *aRight = *paRight;
char *pEnd1 = &aLeft[nLeft];
char *pEnd2 = &aRight[*pnRight];
char *p1 = aLeft;
char *p2 = aRight;
char *p;
int bFirstOut = 0;
char *aOut;
assert( nDist>0 );
if( bDescDoclist ){
aOut = sqlite3_malloc64((sqlite3_int64)*pnRight + FTS3_VARINT_MAX);
if( aOut==0 ) return SQLITE_NOMEM;
}else{
aOut = aRight;
}
p = aOut;
fts3GetDeltaVarint3(&p1, pEnd1, 0, &i1);
fts3GetDeltaVarint3(&p2, pEnd2, 0, &i2);
while( p1 && p2 ){
sqlite3_int64 iDiff = DOCID_CMP(i1, i2);
if( iDiff==0 ){
char *pSave = p;
sqlite3_int64 iPrevSave = iPrev;
int bFirstOutSave = bFirstOut;
fts3PutDeltaVarint3(&p, bDescDoclist, &iPrev, &bFirstOut, i1);
if( 0==fts3PoslistPhraseMerge(&p, nDist, 0, 1, &p1, &p2) ){
p = pSave;
iPrev = iPrevSave;
bFirstOut = bFirstOutSave;
}
fts3GetDeltaVarint3(&p1, pEnd1, bDescDoclist, &i1);
fts3GetDeltaVarint3(&p2, pEnd2, bDescDoclist, &i2);
}else if( iDiff<0 ){
fts3PoslistCopy(0, &p1);
fts3GetDeltaVarint3(&p1, pEnd1, bDescDoclist, &i1);
}else{
fts3PoslistCopy(0, &p2);
fts3GetDeltaVarint3(&p2, pEnd2, bDescDoclist, &i2);
}
}
*pnRight = (int)(p - aOut);
if( bDescDoclist ){
sqlite3_free(aRight);
*paRight = aOut;
}
return SQLITE_OK;
}
/*
** Argument pList points to a position list nList bytes in size. This
** function checks to see if the position list contains any entries for
** a token in position 0 (of any column). If so, it writes argument iDelta
** to the output buffer pOut, followed by a position list consisting only
** of the entries from pList at position 0, and terminated by an 0x00 byte.
** The value returned is the number of bytes written to pOut (if any).
*/
int sqlite3Fts3FirstFilter(
sqlite3_int64 iDelta, /* Varint that may be written to pOut */
char *pList, /* Position list (no 0x00 term) */
int nList, /* Size of pList in bytes */
char *pOut /* Write output here */
){
int nOut = 0;
int bWritten = 0; /* True once iDelta has been written */
char *p = pList;
char *pEnd = &pList[nList];
if( *p!=0x01 ){
if( *p==0x02 ){
nOut += sqlite3Fts3PutVarint(&pOut[nOut], iDelta);
pOut[nOut++] = 0x02;
bWritten = 1;
}
fts3ColumnlistCopy(0, &p);
}
while( p<pEnd ){
sqlite3_int64 iCol;
p++;
p += sqlite3Fts3GetVarint(p, &iCol);
if( *p==0x02 ){
if( bWritten==0 ){
nOut += sqlite3Fts3PutVarint(&pOut[nOut], iDelta);
bWritten = 1;
}
pOut[nOut++] = 0x01;
nOut += sqlite3Fts3PutVarint(&pOut[nOut], iCol);
pOut[nOut++] = 0x02;
}
fts3ColumnlistCopy(0, &p);
}
if( bWritten ){
pOut[nOut++] = 0x00;
}
return nOut;
}
/*
** Merge all doclists in the TermSelect.aaOutput[] array into a single
** doclist stored in TermSelect.aaOutput[0]. If successful, delete all
** other doclists (except the aaOutput[0] one) and return SQLITE_OK.
**
** If an OOM error occurs, return SQLITE_NOMEM. In this case it is
** the responsibility of the caller to free any doclists left in the
** TermSelect.aaOutput[] array.
*/
static int fts3TermSelectFinishMerge(Fts3Table *p, TermSelect *pTS){
char *aOut = 0;
int nOut = 0;
int i;
/* Loop through the doclists in the aaOutput[] array. Merge them all
** into a single doclist.
*/
for(i=0; i<SizeofArray(pTS->aaOutput); i++){
if( pTS->aaOutput[i] ){
if( !aOut ){
aOut = pTS->aaOutput[i];
nOut = pTS->anOutput[i];
pTS->aaOutput[i] = 0;
}else{
int nNew;
char *aNew;
int rc = fts3DoclistOrMerge(p->bDescIdx,
pTS->aaOutput[i], pTS->anOutput[i], aOut, nOut, &aNew, &nNew
);
if( rc!=SQLITE_OK ){
sqlite3_free(aOut);
return rc;
}
sqlite3_free(pTS->aaOutput[i]);
sqlite3_free(aOut);
pTS->aaOutput[i] = 0;
aOut = aNew;
nOut = nNew;
}
}
}
pTS->aaOutput[0] = aOut;
pTS->anOutput[0] = nOut;
return SQLITE_OK;
}
/*
** Merge the doclist aDoclist/nDoclist into the TermSelect object passed
** as the first argument. The merge is an "OR" merge (see function
** fts3DoclistOrMerge() for details).
**
** This function is called with the doclist for each term that matches
** a queried prefix. It merges all these doclists into one, the doclist
** for the specified prefix. Since there can be a very large number of
** doclists to merge, the merging is done pair-wise using the TermSelect
** object.
**
** This function returns SQLITE_OK if the merge is successful, or an
** SQLite error code (SQLITE_NOMEM) if an error occurs.
*/
static int fts3TermSelectMerge(
Fts3Table *p, /* FTS table handle */
TermSelect *pTS, /* TermSelect object to merge into */
char *aDoclist, /* Pointer to doclist */
int nDoclist /* Size of aDoclist in bytes */
){
if( pTS->aaOutput[0]==0 ){
/* If this is the first term selected, copy the doclist to the output
** buffer using memcpy().
**
** Add FTS3_VARINT_MAX bytes of unused space to the end of the
** allocation. This is so as to ensure that the buffer is big enough
** to hold the current doclist AND'd with any other doclist. If the
** doclists are stored in order=ASC order, this padding would not be
** required (since the size of [doclistA AND doclistB] is always less
** than or equal to the size of [doclistA] in that case). But this is
** not true for order=DESC. For example, a doclist containing (1, -1)
** may be smaller than (-1), as in the first example the -1 may be stored
** as a single-byte delta, whereas in the second it must be stored as a
** FTS3_VARINT_MAX byte varint.
**
** Similar padding is added in the fts3DoclistOrMerge() function.
*/
pTS->aaOutput[0] = sqlite3_malloc64((i64)nDoclist + FTS3_VARINT_MAX + 1);
pTS->anOutput[0] = nDoclist;
if( pTS->aaOutput[0] ){
memcpy(pTS->aaOutput[0], aDoclist, nDoclist);
memset(&pTS->aaOutput[0][nDoclist], 0, FTS3_VARINT_MAX);
}else{
return SQLITE_NOMEM;
}
}else{
char *aMerge = aDoclist;
int nMerge = nDoclist;
int iOut;
for(iOut=0; iOut<SizeofArray(pTS->aaOutput); iOut++){
if( pTS->aaOutput[iOut]==0 ){
assert( iOut>0 );
pTS->aaOutput[iOut] = aMerge;
pTS->anOutput[iOut] = nMerge;
break;
}else{
char *aNew;
int nNew;
int rc = fts3DoclistOrMerge(p->bDescIdx, aMerge, nMerge,
pTS->aaOutput[iOut], pTS->anOutput[iOut], &aNew, &nNew
);
if( rc!=SQLITE_OK ){
if( aMerge!=aDoclist ) sqlite3_free(aMerge);
return rc;
}
if( aMerge!=aDoclist ) sqlite3_free(aMerge);
sqlite3_free(pTS->aaOutput[iOut]);
pTS->aaOutput[iOut] = 0;
aMerge = aNew;
nMerge = nNew;
if( (iOut+1)==SizeofArray(pTS->aaOutput) ){
pTS->aaOutput[iOut] = aMerge;
pTS->anOutput[iOut] = nMerge;
}
}
}
}
return SQLITE_OK;
}
/*
** Append SegReader object pNew to the end of the pCsr->apSegment[] array.
*/
static int fts3SegReaderCursorAppend(
Fts3MultiSegReader *pCsr,
Fts3SegReader *pNew
){
if( (pCsr->nSegment%16)==0 ){
Fts3SegReader **apNew;
sqlite3_int64 nByte = (pCsr->nSegment + 16)*sizeof(Fts3SegReader*);
apNew = (Fts3SegReader **)sqlite3_realloc64(pCsr->apSegment, nByte);
if( !apNew ){
sqlite3Fts3SegReaderFree(pNew);
return SQLITE_NOMEM;
}
pCsr->apSegment = apNew;
}
pCsr->apSegment[pCsr->nSegment++] = pNew;
return SQLITE_OK;
}
/*
** Add seg-reader objects to the Fts3MultiSegReader object passed as the
** 8th argument.
**
** This function returns SQLITE_OK if successful, or an SQLite error code
** otherwise.
*/
static int fts3SegReaderCursor(
Fts3Table *p, /* FTS3 table handle */
int iLangid, /* Language id */
int iIndex, /* Index to search (from 0 to p->nIndex-1) */
int iLevel, /* Level of segments to scan */
const char *zTerm, /* Term to query for */
int nTerm, /* Size of zTerm in bytes */
int isPrefix, /* True for a prefix search */
int isScan, /* True to scan from zTerm to EOF */
Fts3MultiSegReader *pCsr /* Cursor object to populate */
){
int rc = SQLITE_OK; /* Error code */
sqlite3_stmt *pStmt = 0; /* Statement to iterate through segments */
int rc2; /* Result of sqlite3_reset() */
/* If iLevel is less than 0 and this is not a scan, include a seg-reader
** for the pending-terms. If this is a scan, then this call must be being
** made by an fts4aux module, not an FTS table. In this case calling
** Fts3SegReaderPending might segfault, as the data structures used by
** fts4aux are not completely populated. So it's easiest to filter these
** calls out here. */
if( iLevel<0 && p->aIndex && p->iPrevLangid==iLangid ){
Fts3SegReader *pSeg = 0;
rc = sqlite3Fts3SegReaderPending(p, iIndex, zTerm, nTerm, isPrefix||isScan, &pSeg);
if( rc==SQLITE_OK && pSeg ){
rc = fts3SegReaderCursorAppend(pCsr, pSeg);
}
}
if( iLevel!=FTS3_SEGCURSOR_PENDING ){
if( rc==SQLITE_OK ){
rc = sqlite3Fts3AllSegdirs(p, iLangid, iIndex, iLevel, &pStmt);
}
while( rc==SQLITE_OK && SQLITE_ROW==(rc = sqlite3_step(pStmt)) ){
Fts3SegReader *pSeg = 0;
/* Read the values returned by the SELECT into local variables. */
sqlite3_int64 iStartBlock = sqlite3_column_int64(pStmt, 1);
sqlite3_int64 iLeavesEndBlock = sqlite3_column_int64(pStmt, 2);
sqlite3_int64 iEndBlock = sqlite3_column_int64(pStmt, 3);
int nRoot = sqlite3_column_bytes(pStmt, 4);
char const *zRoot = sqlite3_column_blob(pStmt, 4);
/* If zTerm is not NULL, and this segment is not stored entirely on its
** root node, the range of leaves scanned can be reduced. Do this. */
if( iStartBlock && zTerm && zRoot ){
sqlite3_int64 *pi = (isPrefix ? &iLeavesEndBlock : 0);
rc = fts3SelectLeaf(p, zTerm, nTerm, zRoot, nRoot, &iStartBlock, pi);
if( rc!=SQLITE_OK ) goto finished;
if( isPrefix==0 && isScan==0 ) iLeavesEndBlock = iStartBlock;
}
rc = sqlite3Fts3SegReaderNew(pCsr->nSegment+1,
(isPrefix==0 && isScan==0),
iStartBlock, iLeavesEndBlock,
iEndBlock, zRoot, nRoot, &pSeg
);
if( rc!=SQLITE_OK ) goto finished;
rc = fts3SegReaderCursorAppend(pCsr, pSeg);
}
}
finished:
rc2 = sqlite3_reset(pStmt);
if( rc==SQLITE_DONE ) rc = rc2;
return rc;
}
/*
** Set up a cursor object for iterating through a full-text index or a
** single level therein.
*/
int sqlite3Fts3SegReaderCursor(
Fts3Table *p, /* FTS3 table handle */
int iLangid, /* Language-id to search */
int iIndex, /* Index to search (from 0 to p->nIndex-1) */
int iLevel, /* Level of segments to scan */
const char *zTerm, /* Term to query for */
int nTerm, /* Size of zTerm in bytes */
int isPrefix, /* True for a prefix search */
int isScan, /* True to scan from zTerm to EOF */
Fts3MultiSegReader *pCsr /* Cursor object to populate */
){
assert( iIndex>=0 && iIndex<p->nIndex );
assert( iLevel==FTS3_SEGCURSOR_ALL
|| iLevel==FTS3_SEGCURSOR_PENDING
|| iLevel>=0
);
assert( iLevel<FTS3_SEGDIR_MAXLEVEL );
assert( FTS3_SEGCURSOR_ALL<0 && FTS3_SEGCURSOR_PENDING<0 );
assert( isPrefix==0 || isScan==0 );
memset(pCsr, 0, sizeof(Fts3MultiSegReader));
return fts3SegReaderCursor(
p, iLangid, iIndex, iLevel, zTerm, nTerm, isPrefix, isScan, pCsr
);
}
/*
** In addition to its current configuration, have the Fts3MultiSegReader
** passed as the 4th argument also scan the doclist for term zTerm/nTerm.
**
** SQLITE_OK is returned if no error occurs, otherwise an SQLite error code.
*/
static int fts3SegReaderCursorAddZero(
Fts3Table *p, /* FTS virtual table handle */
int iLangid,
const char *zTerm, /* Term to scan doclist of */
int nTerm, /* Number of bytes in zTerm */
Fts3MultiSegReader *pCsr /* Fts3MultiSegReader to modify */
){
return fts3SegReaderCursor(p,
iLangid, 0, FTS3_SEGCURSOR_ALL, zTerm, nTerm, 0, 0,pCsr
);
}
/*
** Open an Fts3MultiSegReader to scan the doclist for term zTerm/nTerm. Or,
** if isPrefix is true, to scan the doclist for all terms for which
** zTerm/nTerm is a prefix. If successful, return SQLITE_OK and write
** a pointer to the new Fts3MultiSegReader to *ppSegcsr. Otherwise, return
** an SQLite error code.
**
** It is the responsibility of the caller to free this object by eventually
** passing it to fts3SegReaderCursorFree()
**
** SQLITE_OK is returned if no error occurs, otherwise an SQLite error code.
** Output parameter *ppSegcsr is set to 0 if an error occurs.
*/
static int fts3TermSegReaderCursor(
Fts3Cursor *pCsr, /* Virtual table cursor handle */
const char *zTerm, /* Term to query for */
int nTerm, /* Size of zTerm in bytes */
int isPrefix, /* True for a prefix search */
Fts3MultiSegReader **ppSegcsr /* OUT: Allocated seg-reader cursor */
){
Fts3MultiSegReader *pSegcsr; /* Object to allocate and return */
int rc = SQLITE_NOMEM; /* Return code */
pSegcsr = sqlite3_malloc(sizeof(Fts3MultiSegReader));
if( pSegcsr ){
int i;
int bFound = 0; /* True once an index has been found */
Fts3Table *p = (Fts3Table *)pCsr->base.pVtab;
if( isPrefix ){
for(i=1; bFound==0 && i<p->nIndex; i++){
if( p->aIndex[i].nPrefix==nTerm ){
bFound = 1;
rc = sqlite3Fts3SegReaderCursor(p, pCsr->iLangid,
i, FTS3_SEGCURSOR_ALL, zTerm, nTerm, 0, 0, pSegcsr
);
pSegcsr->bLookup = 1;
}
}
for(i=1; bFound==0 && i<p->nIndex; i++){
if( p->aIndex[i].nPrefix==nTerm+1 ){
bFound = 1;
rc = sqlite3Fts3SegReaderCursor(p, pCsr->iLangid,
i, FTS3_SEGCURSOR_ALL, zTerm, nTerm, 1, 0, pSegcsr
);
if( rc==SQLITE_OK ){
rc = fts3SegReaderCursorAddZero(
p, pCsr->iLangid, zTerm, nTerm, pSegcsr
);
}
}
}
}
if( bFound==0 ){
rc = sqlite3Fts3SegReaderCursor(p, pCsr->iLangid,
0, FTS3_SEGCURSOR_ALL, zTerm, nTerm, isPrefix, 0, pSegcsr
);
pSegcsr->bLookup = !isPrefix;
}
}
*ppSegcsr = pSegcsr;
return rc;
}
/*
** Free an Fts3MultiSegReader allocated by fts3TermSegReaderCursor().
*/
static void fts3SegReaderCursorFree(Fts3MultiSegReader *pSegcsr){
sqlite3Fts3SegReaderFinish(pSegcsr);
sqlite3_free(pSegcsr);
}
/*
** This function retrieves the doclist for the specified term (or term
** prefix) from the database.
*/
static int fts3TermSelect(
Fts3Table *p, /* Virtual table handle */
Fts3PhraseToken *pTok, /* Token to query for */
int iColumn, /* Column to query (or -ve for all columns) */
int *pnOut, /* OUT: Size of buffer at *ppOut */
char **ppOut /* OUT: Malloced result buffer */
){
int rc; /* Return code */
Fts3MultiSegReader *pSegcsr; /* Seg-reader cursor for this term */
TermSelect tsc; /* Object for pair-wise doclist merging */
Fts3SegFilter filter; /* Segment term filter configuration */
pSegcsr = pTok->pSegcsr;
memset(&tsc, 0, sizeof(TermSelect));
filter.flags = FTS3_SEGMENT_IGNORE_EMPTY | FTS3_SEGMENT_REQUIRE_POS
| (pTok->isPrefix ? FTS3_SEGMENT_PREFIX : 0)
| (pTok->bFirst ? FTS3_SEGMENT_FIRST : 0)
| (iColumn<p->nColumn ? FTS3_SEGMENT_COLUMN_FILTER : 0);
filter.iCol = iColumn;
filter.zTerm = pTok->z;
filter.nTerm = pTok->n;
rc = sqlite3Fts3SegReaderStart(p, pSegcsr, &filter);
while( SQLITE_OK==rc
&& SQLITE_ROW==(rc = sqlite3Fts3SegReaderStep(p, pSegcsr))
){
rc = fts3TermSelectMerge(p, &tsc, pSegcsr->aDoclist, pSegcsr->nDoclist);
}
if( rc==SQLITE_OK ){
rc = fts3TermSelectFinishMerge(p, &tsc);
}
if( rc==SQLITE_OK ){
*ppOut = tsc.aaOutput[0];
*pnOut = tsc.anOutput[0];
}else{
int i;
for(i=0; i<SizeofArray(tsc.aaOutput); i++){
sqlite3_free(tsc.aaOutput[i]);
}
}
fts3SegReaderCursorFree(pSegcsr);
pTok->pSegcsr = 0;
return rc;
}
/*
** This function counts the total number of docids in the doclist stored
** in buffer aList[], size nList bytes.
**
** If the isPoslist argument is true, then it is assumed that the doclist
** contains a position-list following each docid. Otherwise, it is assumed
** that the doclist is simply a list of docids stored as delta encoded
** varints.
*/
static int fts3DoclistCountDocids(char *aList, int nList){
int nDoc = 0; /* Return value */
if( aList ){
char *aEnd = &aList[nList]; /* Pointer to one byte after EOF */
char *p = aList; /* Cursor */
while( p<aEnd ){
nDoc++;
while( (*p++)&0x80 ); /* Skip docid varint */
fts3PoslistCopy(0, &p); /* Skip over position list */
}
}
return nDoc;
}
/*
** Advance the cursor to the next row in the %_content table that
** matches the search criteria. For a MATCH search, this will be
** the next row that matches. For a full-table scan, this will be
** simply the next row in the %_content table. For a docid lookup,
** this routine simply sets the EOF flag.
**
** Return SQLITE_OK if nothing goes wrong. SQLITE_OK is returned
** even if we reach end-of-file. The fts3EofMethod() will be called
** subsequently to determine whether or not an EOF was hit.
*/
static int fts3NextMethod(sqlite3_vtab_cursor *pCursor){
int rc;
Fts3Cursor *pCsr = (Fts3Cursor *)pCursor;
if( pCsr->eSearch==FTS3_DOCID_SEARCH || pCsr->eSearch==FTS3_FULLSCAN_SEARCH ){
Fts3Table *pTab = (Fts3Table*)pCursor->pVtab;
pTab->bLock++;
if( SQLITE_ROW!=sqlite3_step(pCsr->pStmt) ){
pCsr->isEof = 1;
rc = sqlite3_reset(pCsr->pStmt);
}else{
pCsr->iPrevId = sqlite3_column_int64(pCsr->pStmt, 0);
rc = SQLITE_OK;
}
pTab->bLock--;
}else{
rc = fts3EvalNext((Fts3Cursor *)pCursor);
}
assert( ((Fts3Table *)pCsr->base.pVtab)->pSegments==0 );
return rc;
}
/*
** If the numeric type of argument pVal is "integer", then return it
** converted to a 64-bit signed integer. Otherwise, return a copy of
** the second parameter, iDefault.
*/
static sqlite3_int64 fts3DocidRange(sqlite3_value *pVal, i64 iDefault){
if( pVal ){
int eType = sqlite3_value_numeric_type(pVal);
if( eType==SQLITE_INTEGER ){
return sqlite3_value_int64(pVal);
}
}
return iDefault;
}
/*
** This is the xFilter interface for the virtual table. See
** the virtual table xFilter method documentation for additional
** information.
**
** If idxNum==FTS3_FULLSCAN_SEARCH then do a full table scan against
** the %_content table.
**
** If idxNum==FTS3_DOCID_SEARCH then do a docid lookup for a single entry
** in the %_content table.
**
** If idxNum>=FTS3_FULLTEXT_SEARCH then use the full text index. The
** column on the left-hand side of the MATCH operator is column
** number idxNum-FTS3_FULLTEXT_SEARCH, 0 indexed. argv[0] is the right-hand
** side of the MATCH operator.
*/
static int fts3FilterMethod(
sqlite3_vtab_cursor *pCursor, /* The cursor used for this query */
int idxNum, /* Strategy index */
const char *idxStr, /* Unused */
int nVal, /* Number of elements in apVal */
sqlite3_value **apVal /* Arguments for the indexing scheme */
){
int rc = SQLITE_OK;
char *zSql; /* SQL statement used to access %_content */
int eSearch;
Fts3Table *p = (Fts3Table *)pCursor->pVtab;
Fts3Cursor *pCsr = (Fts3Cursor *)pCursor;
sqlite3_value *pCons = 0; /* The MATCH or rowid constraint, if any */
sqlite3_value *pLangid = 0; /* The "langid = ?" constraint, if any */
sqlite3_value *pDocidGe = 0; /* The "docid >= ?" constraint, if any */
sqlite3_value *pDocidLe = 0; /* The "docid <= ?" constraint, if any */
int iIdx;
UNUSED_PARAMETER(idxStr);
UNUSED_PARAMETER(nVal);
if( p->bLock ){
return SQLITE_ERROR;
}
eSearch = (idxNum & 0x0000FFFF);
assert( eSearch>=0 && eSearch<=(FTS3_FULLTEXT_SEARCH+p->nColumn) );
assert( p->pSegments==0 );
/* Collect arguments into local variables */
iIdx = 0;
if( eSearch!=FTS3_FULLSCAN_SEARCH ) pCons = apVal[iIdx++];
if( idxNum & FTS3_HAVE_LANGID ) pLangid = apVal[iIdx++];
if( idxNum & FTS3_HAVE_DOCID_GE ) pDocidGe = apVal[iIdx++];
if( idxNum & FTS3_HAVE_DOCID_LE ) pDocidLe = apVal[iIdx++];
assert( iIdx==nVal );
/* In case the cursor has been used before, clear it now. */
fts3ClearCursor(pCsr);
/* Set the lower and upper bounds on docids to return */
pCsr->iMinDocid = fts3DocidRange(pDocidGe, SMALLEST_INT64);
pCsr->iMaxDocid = fts3DocidRange(pDocidLe, LARGEST_INT64);
if( idxStr ){
pCsr->bDesc = (idxStr[0]=='D');
}else{
pCsr->bDesc = p->bDescIdx;
}
pCsr->eSearch = (i16)eSearch;
if( eSearch!=FTS3_DOCID_SEARCH && eSearch!=FTS3_FULLSCAN_SEARCH ){
int iCol = eSearch-FTS3_FULLTEXT_SEARCH;
const char *zQuery = (const char *)sqlite3_value_text(pCons);
if( zQuery==0 && sqlite3_value_type(pCons)!=SQLITE_NULL ){
return SQLITE_NOMEM;
}
pCsr->iLangid = 0;
if( pLangid ) pCsr->iLangid = sqlite3_value_int(pLangid);
assert( p->base.zErrMsg==0 );
rc = sqlite3Fts3ExprParse(p->pTokenizer, pCsr->iLangid,
p->azColumn, p->bFts4, p->nColumn, iCol, zQuery, -1, &pCsr->pExpr,
&p->base.zErrMsg
);
if( rc!=SQLITE_OK ){
return rc;
}
rc = fts3EvalStart(pCsr);
sqlite3Fts3SegmentsClose(p);
if( rc!=SQLITE_OK ) return rc;
pCsr->pNextId = pCsr->aDoclist;
pCsr->iPrevId = 0;
}
/* Compile a SELECT statement for this cursor. For a full-table-scan, the
** statement loops through all rows of the %_content table. For a
** full-text query or docid lookup, the statement retrieves a single
** row by docid.
*/
if( eSearch==FTS3_FULLSCAN_SEARCH ){
if( pDocidGe || pDocidLe ){
zSql = sqlite3_mprintf(
"SELECT %s WHERE rowid BETWEEN %lld AND %lld ORDER BY rowid %s",
p->zReadExprlist, pCsr->iMinDocid, pCsr->iMaxDocid,
(pCsr->bDesc ? "DESC" : "ASC")
);
}else{
zSql = sqlite3_mprintf("SELECT %s ORDER BY rowid %s",
p->zReadExprlist, (pCsr->bDesc ? "DESC" : "ASC")
);
}
if( zSql ){
p->bLock++;
rc = sqlite3_prepare_v3(
p->db,zSql,-1,SQLITE_PREPARE_PERSISTENT,&pCsr->pStmt,0
);
p->bLock--;
sqlite3_free(zSql);
}else{
rc = SQLITE_NOMEM;
}
}else if( eSearch==FTS3_DOCID_SEARCH ){
rc = fts3CursorSeekStmt(pCsr);
if( rc==SQLITE_OK ){
rc = sqlite3_bind_value(pCsr->pStmt, 1, pCons);
}
}
if( rc!=SQLITE_OK ) return rc;
return fts3NextMethod(pCursor);
}
/*
** This is the xEof method of the virtual table. SQLite calls this
** routine to find out if it has reached the end of a result set.
*/
static int fts3EofMethod(sqlite3_vtab_cursor *pCursor){
Fts3Cursor *pCsr = (Fts3Cursor*)pCursor;
if( pCsr->isEof ){
fts3ClearCursor(pCsr);
pCsr->isEof = 1;
}
return pCsr->isEof;
}
/*
** This is the xRowid method. The SQLite core calls this routine to
** retrieve the rowid for the current row of the result set. fts3
** exposes %_content.docid as the rowid for the virtual table. The
** rowid should be written to *pRowid.
*/
static int fts3RowidMethod(sqlite3_vtab_cursor *pCursor, sqlite_int64 *pRowid){
Fts3Cursor *pCsr = (Fts3Cursor *) pCursor;
*pRowid = pCsr->iPrevId;
return SQLITE_OK;
}
/*
** This is the xColumn method, called by SQLite to request a value from
** the row that the supplied cursor currently points to.
**
** If:
**
** (iCol < p->nColumn) -> The value of the iCol'th user column.
** (iCol == p->nColumn) -> Magic column with the same name as the table.
** (iCol == p->nColumn+1) -> Docid column
** (iCol == p->nColumn+2) -> Langid column
*/
static int fts3ColumnMethod(
sqlite3_vtab_cursor *pCursor, /* Cursor to retrieve value from */
sqlite3_context *pCtx, /* Context for sqlite3_result_xxx() calls */
int iCol /* Index of column to read value from */
){
int rc = SQLITE_OK; /* Return Code */
Fts3Cursor *pCsr = (Fts3Cursor *) pCursor;
Fts3Table *p = (Fts3Table *)pCursor->pVtab;
/* The column value supplied by SQLite must be in range. */
assert( iCol>=0 && iCol<=p->nColumn+2 );
switch( iCol-p->nColumn ){
case 0:
/* The special 'table-name' column */
sqlite3_result_pointer(pCtx, pCsr, "fts3cursor", 0);
break;
case 1:
/* The docid column */
sqlite3_result_int64(pCtx, pCsr->iPrevId);
break;
case 2:
if( pCsr->pExpr ){
sqlite3_result_int64(pCtx, pCsr->iLangid);
break;
}else if( p->zLanguageid==0 ){
sqlite3_result_int(pCtx, 0);
break;
}else{
iCol = p->nColumn;
/* no break */ deliberate_fall_through
}
default:
/* A user column. Or, if this is a full-table scan, possibly the
** language-id column. Seek the cursor. */
rc = fts3CursorSeek(0, pCsr);
if( rc==SQLITE_OK && sqlite3_data_count(pCsr->pStmt)-1>iCol ){
sqlite3_result_value(pCtx, sqlite3_column_value(pCsr->pStmt, iCol+1));
}
break;
}
assert( ((Fts3Table *)pCsr->base.pVtab)->pSegments==0 );
return rc;
}
/*
** This function is the implementation of the xUpdate callback used by
** FTS3 virtual tables. It is invoked by SQLite each time a row is to be
** inserted, updated or deleted.
*/
static int fts3UpdateMethod(
sqlite3_vtab *pVtab, /* Virtual table handle */
int nArg, /* Size of argument array */
sqlite3_value **apVal, /* Array of arguments */
sqlite_int64 *pRowid /* OUT: The affected (or effected) rowid */
){
return sqlite3Fts3UpdateMethod(pVtab, nArg, apVal, pRowid);
}
/*
** Implementation of xSync() method. Flush the contents of the pending-terms
** hash-table to the database.
*/
static int fts3SyncMethod(sqlite3_vtab *pVtab){
/* Following an incremental-merge operation, assuming that the input
** segments are not completely consumed (the usual case), they are updated
** in place to remove the entries that have already been merged. This
** involves updating the leaf block that contains the smallest unmerged
** entry and each block (if any) between the leaf and the root node. So
** if the height of the input segment b-trees is N, and input segments
** are merged eight at a time, updating the input segments at the end
** of an incremental-merge requires writing (8*(1+N)) blocks. N is usually
** small - often between 0 and 2. So the overhead of the incremental
** merge is somewhere between 8 and 24 blocks. To avoid this overhead
** dwarfing the actual productive work accomplished, the incremental merge
** is only attempted if it will write at least 64 leaf blocks. Hence
** nMinMerge.
**
** Of course, updating the input segments also involves deleting a bunch
** of blocks from the segments table. But this is not considered overhead
** as it would also be required by a crisis-merge that used the same input
** segments.
*/
const u32 nMinMerge = 64; /* Minimum amount of incr-merge work to do */
Fts3Table *p = (Fts3Table*)pVtab;
int rc;
i64 iLastRowid = sqlite3_last_insert_rowid(p->db);
rc = sqlite3Fts3PendingTermsFlush(p);
if( rc==SQLITE_OK
&& p->nLeafAdd>(nMinMerge/16)
&& p->nAutoincrmerge && p->nAutoincrmerge!=0xff
){
int mxLevel = 0; /* Maximum relative level value in db */
int A; /* Incr-merge parameter A */
rc = sqlite3Fts3MaxLevel(p, &mxLevel);
assert( rc==SQLITE_OK || mxLevel==0 );
A = p->nLeafAdd * mxLevel;
A += (A/2);
if( A>(int)nMinMerge ) rc = sqlite3Fts3Incrmerge(p, A, p->nAutoincrmerge);
}
sqlite3Fts3SegmentsClose(p);
sqlite3_set_last_insert_rowid(p->db, iLastRowid);
return rc;
}
/*
** If it is currently unknown whether or not the FTS table has an %_stat
** table (if p->bHasStat==2), attempt to determine this (set p->bHasStat
** to 0 or 1). Return SQLITE_OK if successful, or an SQLite error code
** if an error occurs.
*/
static int fts3SetHasStat(Fts3Table *p){
int rc = SQLITE_OK;
if( p->bHasStat==2 ){
char *zTbl = sqlite3_mprintf("%s_stat", p->zName);
if( zTbl ){
int res = sqlite3_table_column_metadata(p->db, p->zDb, zTbl, 0,0,0,0,0,0);
sqlite3_free(zTbl);
p->bHasStat = (res==SQLITE_OK);
}else{
rc = SQLITE_NOMEM;
}
}
return rc;
}
/*
** Implementation of xBegin() method.
*/
static int fts3BeginMethod(sqlite3_vtab *pVtab){
Fts3Table *p = (Fts3Table*)pVtab;
int rc;
UNUSED_PARAMETER(pVtab);
assert( p->pSegments==0 );
assert( p->nPendingData==0 );
assert( p->inTransaction!=1 );
p->nLeafAdd = 0;
rc = fts3SetHasStat(p);
#ifdef SQLITE_DEBUG
if( rc==SQLITE_OK ){
p->inTransaction = 1;
p->mxSavepoint = -1;
}
#endif
return rc;
}
/*
** Implementation of xCommit() method. This is a no-op. The contents of
** the pending-terms hash-table have already been flushed into the database
** by fts3SyncMethod().
*/
static int fts3CommitMethod(sqlite3_vtab *pVtab){
TESTONLY( Fts3Table *p = (Fts3Table*)pVtab );
UNUSED_PARAMETER(pVtab);
assert( p->nPendingData==0 );
assert( p->inTransaction!=0 );
assert( p->pSegments==0 );
TESTONLY( p->inTransaction = 0 );
TESTONLY( p->mxSavepoint = -1; );
return SQLITE_OK;
}
/*
** Implementation of xRollback(). Discard the contents of the pending-terms
** hash-table. Any changes made to the database are reverted by SQLite.
*/
static int fts3RollbackMethod(sqlite3_vtab *pVtab){
Fts3Table *p = (Fts3Table*)pVtab;
sqlite3Fts3PendingTermsClear(p);
assert( p->inTransaction!=0 );
TESTONLY( p->inTransaction = 0 );
TESTONLY( p->mxSavepoint = -1; );
return SQLITE_OK;
}
/*
** When called, *ppPoslist must point to the byte immediately following the
** end of a position-list. i.e. ( (*ppPoslist)[-1]==POS_END ). This function
** moves *ppPoslist so that it instead points to the first byte of the
** same position list.
*/
static void fts3ReversePoslist(char *pStart, char **ppPoslist){
char *p = &(*ppPoslist)[-2];
char c = 0;
/* Skip backwards passed any trailing 0x00 bytes added by NearTrim() */
while( p>pStart && (c=*p--)==0 );
/* Search backwards for a varint with value zero (the end of the previous
** poslist). This is an 0x00 byte preceded by some byte that does not
** have the 0x80 bit set. */
while( p>pStart && (*p & 0x80) | c ){
c = *p--;
}
assert( p==pStart || c==0 );
/* At this point p points to that preceding byte without the 0x80 bit
** set. So to find the start of the poslist, skip forward 2 bytes then
** over a varint.
**
** Normally. The other case is that p==pStart and the poslist to return
** is the first in the doclist. In this case do not skip forward 2 bytes.
** The second part of the if condition (c==0 && *ppPoslist>&p[2])
** is required for cases where the first byte of a doclist and the
** doclist is empty. For example, if the first docid is 10, a doclist
** that begins with:
**
** 0x0A 0x00 <next docid delta varint>
*/
if( p>pStart || (c==0 && *ppPoslist>&p[2]) ){ p = &p[2]; }
while( *p++&0x80 );
*ppPoslist = p;
}
/*
** Helper function used by the implementation of the overloaded snippet(),
** offsets() and optimize() SQL functions.
**
** If the value passed as the third argument is a blob of size
** sizeof(Fts3Cursor*), then the blob contents are copied to the
** output variable *ppCsr and SQLITE_OK is returned. Otherwise, an error
** message is written to context pContext and SQLITE_ERROR returned. The
** string passed via zFunc is used as part of the error message.
*/
static int fts3FunctionArg(
sqlite3_context *pContext, /* SQL function call context */
const char *zFunc, /* Function name */
sqlite3_value *pVal, /* argv[0] passed to function */
Fts3Cursor **ppCsr /* OUT: Store cursor handle here */
){
int rc;
*ppCsr = (Fts3Cursor*)sqlite3_value_pointer(pVal, "fts3cursor");
if( (*ppCsr)!=0 ){
rc = SQLITE_OK;
}else{
char *zErr = sqlite3_mprintf("illegal first argument to %s", zFunc);
sqlite3_result_error(pContext, zErr, -1);
sqlite3_free(zErr);
rc = SQLITE_ERROR;
}
return rc;
}
/*
** Implementation of the snippet() function for FTS3
*/
static void fts3SnippetFunc(
sqlite3_context *pContext, /* SQLite function call context */
int nVal, /* Size of apVal[] array */
sqlite3_value **apVal /* Array of arguments */
){
Fts3Cursor *pCsr; /* Cursor handle passed through apVal[0] */
const char *zStart = "<b>";
const char *zEnd = "</b>";
const char *zEllipsis = "<b>...</b>";
int iCol = -1;
int nToken = 15; /* Default number of tokens in snippet */
/* There must be at least one argument passed to this function (otherwise
** the non-overloaded version would have been called instead of this one).
*/
assert( nVal>=1 );
if( nVal>6 ){
sqlite3_result_error(pContext,
"wrong number of arguments to function snippet()", -1);
return;
}
if( fts3FunctionArg(pContext, "snippet", apVal[0], &pCsr) ) return;
switch( nVal ){
case 6: nToken = sqlite3_value_int(apVal[5]);
/* no break */ deliberate_fall_through
case 5: iCol = sqlite3_value_int(apVal[4]);
/* no break */ deliberate_fall_through
case 4: zEllipsis = (const char*)sqlite3_value_text(apVal[3]);
/* no break */ deliberate_fall_through
case 3: zEnd = (const char*)sqlite3_value_text(apVal[2]);
/* no break */ deliberate_fall_through
case 2: zStart = (const char*)sqlite3_value_text(apVal[1]);
}
if( !zEllipsis || !zEnd || !zStart ){
sqlite3_result_error_nomem(pContext);
}else if( nToken==0 ){
sqlite3_result_text(pContext, "", -1, SQLITE_STATIC);
}else if( SQLITE_OK==fts3CursorSeek(pContext, pCsr) ){
sqlite3Fts3Snippet(pContext, pCsr, zStart, zEnd, zEllipsis, iCol, nToken);
}
}
/*
** Implementation of the offsets() function for FTS3
*/
static void fts3OffsetsFunc(
sqlite3_context *pContext, /* SQLite function call context */
int nVal, /* Size of argument array */
sqlite3_value **apVal /* Array of arguments */
){
Fts3Cursor *pCsr; /* Cursor handle passed through apVal[0] */
UNUSED_PARAMETER(nVal);
assert( nVal==1 );
if( fts3FunctionArg(pContext, "offsets", apVal[0], &pCsr) ) return;
assert( pCsr );
if( SQLITE_OK==fts3CursorSeek(pContext, pCsr) ){
sqlite3Fts3Offsets(pContext, pCsr);
}
}
/*
** Implementation of the special optimize() function for FTS3. This
** function merges all segments in the database to a single segment.
** Example usage is:
**
** SELECT optimize(t) FROM t LIMIT 1;
**
** where 't' is the name of an FTS3 table.
*/
static void fts3OptimizeFunc(
sqlite3_context *pContext, /* SQLite function call context */
int nVal, /* Size of argument array */
sqlite3_value **apVal /* Array of arguments */
){
int rc; /* Return code */
Fts3Table *p; /* Virtual table handle */
Fts3Cursor *pCursor; /* Cursor handle passed through apVal[0] */
UNUSED_PARAMETER(nVal);
assert( nVal==1 );
if( fts3FunctionArg(pContext, "optimize", apVal[0], &pCursor) ) return;
p = (Fts3Table *)pCursor->base.pVtab;
assert( p );
rc = sqlite3Fts3Optimize(p);
switch( rc ){
case SQLITE_OK:
sqlite3_result_text(pContext, "Index optimized", -1, SQLITE_STATIC);
break;
case SQLITE_DONE:
sqlite3_result_text(pContext, "Index already optimal", -1, SQLITE_STATIC);
break;
default:
sqlite3_result_error_code(pContext, rc);
break;
}
}
/*
** Implementation of the matchinfo() function for FTS3
*/
static void fts3MatchinfoFunc(
sqlite3_context *pContext, /* SQLite function call context */
int nVal, /* Size of argument array */
sqlite3_value **apVal /* Array of arguments */
){
Fts3Cursor *pCsr; /* Cursor handle passed through apVal[0] */
assert( nVal==1 || nVal==2 );
if( SQLITE_OK==fts3FunctionArg(pContext, "matchinfo", apVal[0], &pCsr) ){
const char *zArg = 0;
if( nVal>1 ){
zArg = (const char *)sqlite3_value_text(apVal[1]);
}
sqlite3Fts3Matchinfo(pContext, pCsr, zArg);
}
}
/*
** This routine implements the xFindFunction method for the FTS3
** virtual table.
*/
static int fts3FindFunctionMethod(
sqlite3_vtab *pVtab, /* Virtual table handle */
int nArg, /* Number of SQL function arguments */
const char *zName, /* Name of SQL function */
void (**pxFunc)(sqlite3_context*,int,sqlite3_value**), /* OUT: Result */
void **ppArg /* Unused */
){
struct Overloaded {
const char *zName;
void (*xFunc)(sqlite3_context*,int,sqlite3_value**);
} aOverload[] = {
{ "snippet", fts3SnippetFunc },
{ "offsets", fts3OffsetsFunc },
{ "optimize", fts3OptimizeFunc },
{ "matchinfo", fts3MatchinfoFunc },
};
int i; /* Iterator variable */
UNUSED_PARAMETER(pVtab);
UNUSED_PARAMETER(nArg);
UNUSED_PARAMETER(ppArg);
for(i=0; i<SizeofArray(aOverload); i++){
if( strcmp(zName, aOverload[i].zName)==0 ){
*pxFunc = aOverload[i].xFunc;
return 1;
}
}
/* No function of the specified name was found. Return 0. */
return 0;
}
/*
** Implementation of FTS3 xRename method. Rename an fts3 table.
*/
static int fts3RenameMethod(
sqlite3_vtab *pVtab, /* Virtual table handle */
const char *zName /* New name of table */
){
Fts3Table *p = (Fts3Table *)pVtab;
sqlite3 *db = p->db; /* Database connection */
int rc; /* Return Code */
/* At this point it must be known if the %_stat table exists or not.
** So bHasStat may not be 2. */
rc = fts3SetHasStat(p);
/* As it happens, the pending terms table is always empty here. This is
** because an "ALTER TABLE RENAME TABLE" statement inside a transaction
** always opens a savepoint transaction. And the xSavepoint() method
** flushes the pending terms table. But leave the (no-op) call to
** PendingTermsFlush() in in case that changes.
*/
assert( p->nPendingData==0 );
if( rc==SQLITE_OK ){
rc = sqlite3Fts3PendingTermsFlush(p);
}
p->bIgnoreSavepoint = 1;
if( p->zContentTbl==0 ){
fts3DbExec(&rc, db,
"ALTER TABLE %Q.'%q_content' RENAME TO '%q_content';",
p->zDb, p->zName, zName
);
}
if( p->bHasDocsize ){
fts3DbExec(&rc, db,
"ALTER TABLE %Q.'%q_docsize' RENAME TO '%q_docsize';",
p->zDb, p->zName, zName
);
}
if( p->bHasStat ){
fts3DbExec(&rc, db,
"ALTER TABLE %Q.'%q_stat' RENAME TO '%q_stat';",
p->zDb, p->zName, zName
);
}
fts3DbExec(&rc, db,
"ALTER TABLE %Q.'%q_segments' RENAME TO '%q_segments';",
p->zDb, p->zName, zName
);
fts3DbExec(&rc, db,
"ALTER TABLE %Q.'%q_segdir' RENAME TO '%q_segdir';",
p->zDb, p->zName, zName
);
p->bIgnoreSavepoint = 0;
return rc;
}
/*
** The xSavepoint() method.
**
** Flush the contents of the pending-terms table to disk.
*/
static int fts3SavepointMethod(sqlite3_vtab *pVtab, int iSavepoint){
int rc = SQLITE_OK;
Fts3Table *pTab = (Fts3Table*)pVtab;
assert( pTab->inTransaction );
assert( pTab->mxSavepoint<=iSavepoint );
TESTONLY( pTab->mxSavepoint = iSavepoint );
if( pTab->bIgnoreSavepoint==0 ){
if( fts3HashCount(&pTab->aIndex[0].hPending)>0 ){
char *zSql = sqlite3_mprintf("INSERT INTO %Q.%Q(%Q) VALUES('flush')",
pTab->zDb, pTab->zName, pTab->zName
);
if( zSql ){
pTab->bIgnoreSavepoint = 1;
rc = sqlite3_exec(pTab->db, zSql, 0, 0, 0);
pTab->bIgnoreSavepoint = 0;
sqlite3_free(zSql);
}else{
rc = SQLITE_NOMEM;
}
}
if( rc==SQLITE_OK ){
pTab->iSavepoint = iSavepoint+1;
}
}
return rc;
}
/*
** The xRelease() method.
**
** This is a no-op.
*/
static int fts3ReleaseMethod(sqlite3_vtab *pVtab, int iSavepoint){
Fts3Table *pTab = (Fts3Table*)pVtab;
assert( pTab->inTransaction );
assert( pTab->mxSavepoint >= iSavepoint );
TESTONLY( pTab->mxSavepoint = iSavepoint-1 );
pTab->iSavepoint = iSavepoint;
return SQLITE_OK;
}
/*
** The xRollbackTo() method.
**
** Discard the contents of the pending terms table.
*/
static int fts3RollbackToMethod(sqlite3_vtab *pVtab, int iSavepoint){
Fts3Table *pTab = (Fts3Table*)pVtab;
UNUSED_PARAMETER(iSavepoint);
assert( pTab->inTransaction );
TESTONLY( pTab->mxSavepoint = iSavepoint );
if( (iSavepoint+1)<=pTab->iSavepoint ){
sqlite3Fts3PendingTermsClear(pTab);
}
return SQLITE_OK;
}
/*
** Return true if zName is the extension on one of the shadow tables used
** by this module.
*/
static int fts3ShadowName(const char *zName){
static const char *azName[] = {
"content", "docsize", "segdir", "segments", "stat",
};
unsigned int i;
for(i=0; i<sizeof(azName)/sizeof(azName[0]); i++){
if( sqlite3_stricmp(zName, azName[i])==0 ) return 1;
}
return 0;
}
/*
** Implementation of the xIntegrity() method on the FTS3/FTS4 virtual
** table.
*/
static int fts3IntegrityMethod(
sqlite3_vtab *pVtab, /* The virtual table to be checked */
const char *zSchema, /* Name of schema in which pVtab lives */
const char *zTabname, /* Name of the pVTab table */
int isQuick, /* True if this is a quick_check */
char **pzErr /* Write error message here */
){
Fts3Table *p = (Fts3Table*)pVtab;
int rc = SQLITE_OK;
int bOk = 0;
UNUSED_PARAMETER(isQuick);
rc = sqlite3Fts3IntegrityCheck(p, &bOk);
assert( rc!=SQLITE_CORRUPT_VTAB );
if( rc==SQLITE_ERROR || (rc&0xFF)==SQLITE_CORRUPT ){
*pzErr = sqlite3_mprintf("unable to validate the inverted index for"
" FTS%d table %s.%s: %s",
p->bFts4 ? 4 : 3, zSchema, zTabname, sqlite3_errstr(rc));
if( *pzErr ) rc = SQLITE_OK;
}else if( rc==SQLITE_OK && bOk==0 ){
*pzErr = sqlite3_mprintf("malformed inverted index for FTS%d table %s.%s",
p->bFts4 ? 4 : 3, zSchema, zTabname);
if( *pzErr==0 ) rc = SQLITE_NOMEM;
}
sqlite3Fts3SegmentsClose(p);
return rc;
}
static const sqlite3_module fts3Module = {
/* iVersion */ 4,
/* xCreate */ fts3CreateMethod,
/* xConnect */ fts3ConnectMethod,
/* xBestIndex */ fts3BestIndexMethod,
/* xDisconnect */ fts3DisconnectMethod,
/* xDestroy */ fts3DestroyMethod,
/* xOpen */ fts3OpenMethod,
/* xClose */ fts3CloseMethod,
/* xFilter */ fts3FilterMethod,
/* xNext */ fts3NextMethod,
/* xEof */ fts3EofMethod,
/* xColumn */ fts3ColumnMethod,
/* xRowid */ fts3RowidMethod,
/* xUpdate */ fts3UpdateMethod,
/* xBegin */ fts3BeginMethod,
/* xSync */ fts3SyncMethod,
/* xCommit */ fts3CommitMethod,
/* xRollback */ fts3RollbackMethod,
/* xFindFunction */ fts3FindFunctionMethod,
/* xRename */ fts3RenameMethod,
/* xSavepoint */ fts3SavepointMethod,
/* xRelease */ fts3ReleaseMethod,
/* xRollbackTo */ fts3RollbackToMethod,
/* xShadowName */ fts3ShadowName,
/* xIntegrity */ fts3IntegrityMethod,
};
/*
** This function is registered as the module destructor (called when an
** FTS3 enabled database connection is closed). It frees the memory
** allocated for the tokenizer hash table.
*/
static void hashDestroy(void *p){
Fts3HashWrapper *pHash = (Fts3HashWrapper *)p;
pHash->nRef--;
if( pHash->nRef<=0 ){
sqlite3Fts3HashClear(&pHash->hash);
sqlite3_free(pHash);
}
}
/*
** The fts3 built-in tokenizers - "simple", "porter" and "icu"- are
** implemented in files fts3_tokenizer1.c, fts3_porter.c and fts3_icu.c
** respectively. The following three forward declarations are for functions
** declared in these files used to retrieve the respective implementations.
**
** Calling sqlite3Fts3SimpleTokenizerModule() sets the value pointed
** to by the argument to point to the "simple" tokenizer implementation.
** And so on.
*/
void sqlite3Fts3SimpleTokenizerModule(sqlite3_tokenizer_module const**ppModule);
void sqlite3Fts3PorterTokenizerModule(sqlite3_tokenizer_module const**ppModule);
#ifndef SQLITE_DISABLE_FTS3_UNICODE
void sqlite3Fts3UnicodeTokenizer(sqlite3_tokenizer_module const**ppModule);
#endif
#ifdef SQLITE_ENABLE_ICU
void sqlite3Fts3IcuTokenizerModule(sqlite3_tokenizer_module const**ppModule);
#endif
/*
** Initialize the fts3 extension. If this extension is built as part
** of the sqlite library, then this function is called directly by
** SQLite. If fts3 is built as a dynamically loadable extension, this
** function is called by the sqlite3_extension_init() entry point.
*/
int sqlite3Fts3Init(sqlite3 *db){
int rc = SQLITE_OK;
Fts3HashWrapper *pHash = 0;
const sqlite3_tokenizer_module *pSimple = 0;
const sqlite3_tokenizer_module *pPorter = 0;
#ifndef SQLITE_DISABLE_FTS3_UNICODE
const sqlite3_tokenizer_module *pUnicode = 0;
#endif
#ifdef SQLITE_ENABLE_ICU
const sqlite3_tokenizer_module *pIcu = 0;
sqlite3Fts3IcuTokenizerModule(&pIcu);
#endif
#ifndef SQLITE_DISABLE_FTS3_UNICODE
sqlite3Fts3UnicodeTokenizer(&pUnicode);
#endif
#ifdef SQLITE_TEST
rc = sqlite3Fts3InitTerm(db);
if( rc!=SQLITE_OK ) return rc;
#endif
rc = sqlite3Fts3InitAux(db);
if( rc!=SQLITE_OK ) return rc;
sqlite3Fts3SimpleTokenizerModule(&pSimple);
sqlite3Fts3PorterTokenizerModule(&pPorter);
/* Allocate and initialize the hash-table used to store tokenizers. */
pHash = sqlite3_malloc(sizeof(Fts3HashWrapper));
if( !pHash ){
rc = SQLITE_NOMEM;
}else{
sqlite3Fts3HashInit(&pHash->hash, FTS3_HASH_STRING, 1);
pHash->nRef = 0;
}
/* Load the built-in tokenizers into the hash table */
if( rc==SQLITE_OK ){
if( sqlite3Fts3HashInsert(&pHash->hash, "simple", 7, (void *)pSimple)
|| sqlite3Fts3HashInsert(&pHash->hash, "porter", 7, (void *)pPorter)
#ifndef SQLITE_DISABLE_FTS3_UNICODE
|| sqlite3Fts3HashInsert(&pHash->hash, "unicode61", 10, (void *)pUnicode)
#endif
#ifdef SQLITE_ENABLE_ICU
|| (pIcu && sqlite3Fts3HashInsert(&pHash->hash, "icu", 4, (void *)pIcu))
#endif
){
rc = SQLITE_NOMEM;
}
}
#ifdef SQLITE_TEST
if( rc==SQLITE_OK ){
rc = sqlite3Fts3ExprInitTestInterface(db, &pHash->hash);
}
#endif
/* Create the virtual table wrapper around the hash-table and overload
** the four scalar functions. If this is successful, register the
** module with sqlite.
*/
if( SQLITE_OK==rc
&& SQLITE_OK==(rc=sqlite3Fts3InitHashTable(db,&pHash->hash,"fts3_tokenizer"))
&& SQLITE_OK==(rc = sqlite3_overload_function(db, "snippet", -1))
&& SQLITE_OK==(rc = sqlite3_overload_function(db, "offsets", 1))
&& SQLITE_OK==(rc = sqlite3_overload_function(db, "matchinfo", 1))
&& SQLITE_OK==(rc = sqlite3_overload_function(db, "matchinfo", 2))
&& SQLITE_OK==(rc = sqlite3_overload_function(db, "optimize", 1))
){
pHash->nRef++;
rc = sqlite3_create_module_v2(
db, "fts3", &fts3Module, (void *)pHash, hashDestroy
);
if( rc==SQLITE_OK ){
pHash->nRef++;
rc = sqlite3_create_module_v2(
db, "fts4", &fts3Module, (void *)pHash, hashDestroy
);
}
if( rc==SQLITE_OK ){
pHash->nRef++;
rc = sqlite3Fts3InitTok(db, (void *)pHash, hashDestroy);
}
return rc;
}
/* An error has occurred. Delete the hash table and return the error code. */
assert( rc!=SQLITE_OK );
if( pHash ){
sqlite3Fts3HashClear(&pHash->hash);
sqlite3_free(pHash);
}
return rc;
}
/*
** Allocate an Fts3MultiSegReader for each token in the expression headed
** by pExpr.
**
** An Fts3SegReader object is a cursor that can seek or scan a range of
** entries within a single segment b-tree. An Fts3MultiSegReader uses multiple
** Fts3SegReader objects internally to provide an interface to seek or scan
** within the union of all segments of a b-tree. Hence the name.
**
** If the allocated Fts3MultiSegReader just seeks to a single entry in a
** segment b-tree (if the term is not a prefix or it is a prefix for which
** there exists prefix b-tree of the right length) then it may be traversed
** and merged incrementally. Otherwise, it has to be merged into an in-memory
** doclist and then traversed.
*/
static void fts3EvalAllocateReaders(
Fts3Cursor *pCsr, /* FTS cursor handle */
Fts3Expr *pExpr, /* Allocate readers for this expression */
int *pnToken, /* OUT: Total number of tokens in phrase. */
int *pnOr, /* OUT: Total number of OR nodes in expr. */
int *pRc /* IN/OUT: Error code */
){
if( pExpr && SQLITE_OK==*pRc ){
if( pExpr->eType==FTSQUERY_PHRASE ){
int i;
int nToken = pExpr->pPhrase->nToken;
*pnToken += nToken;
for(i=0; i<nToken; i++){
Fts3PhraseToken *pToken = &pExpr->pPhrase->aToken[i];
int rc = fts3TermSegReaderCursor(pCsr,
pToken->z, pToken->n, pToken->isPrefix, &pToken->pSegcsr
);
if( rc!=SQLITE_OK ){
*pRc = rc;
return;
}
}
assert( pExpr->pPhrase->iDoclistToken==0 );
pExpr->pPhrase->iDoclistToken = -1;
}else{
*pnOr += (pExpr->eType==FTSQUERY_OR);
fts3EvalAllocateReaders(pCsr, pExpr->pLeft, pnToken, pnOr, pRc);
fts3EvalAllocateReaders(pCsr, pExpr->pRight, pnToken, pnOr, pRc);
}
}
}
/*
** Arguments pList/nList contain the doclist for token iToken of phrase p.
** It is merged into the main doclist stored in p->doclist.aAll/nAll.
**
** This function assumes that pList points to a buffer allocated using
** sqlite3_malloc(). This function takes responsibility for eventually
** freeing the buffer.
**
** SQLITE_OK is returned if successful, or SQLITE_NOMEM if an error occurs.
*/
static int fts3EvalPhraseMergeToken(
Fts3Table *pTab, /* FTS Table pointer */
Fts3Phrase *p, /* Phrase to merge pList/nList into */
int iToken, /* Token pList/nList corresponds to */
char *pList, /* Pointer to doclist */
int nList /* Number of bytes in pList */
){
int rc = SQLITE_OK;
assert( iToken!=p->iDoclistToken );
if( pList==0 ){
sqlite3_free(p->doclist.aAll);
p->doclist.aAll = 0;
p->doclist.nAll = 0;
}
else if( p->iDoclistToken<0 ){
p->doclist.aAll = pList;
p->doclist.nAll = nList;
}
else if( p->doclist.aAll==0 ){
sqlite3_free(pList);
}
else {
char *pLeft;
char *pRight;
int nLeft;
int nRight;
int nDiff;
if( p->iDoclistToken<iToken ){
pLeft = p->doclist.aAll;
nLeft = p->doclist.nAll;
pRight = pList;
nRight = nList;
nDiff = iToken - p->iDoclistToken;
}else{
pRight = p->doclist.aAll;
nRight = p->doclist.nAll;
pLeft = pList;
nLeft = nList;
nDiff = p->iDoclistToken - iToken;
}
rc = fts3DoclistPhraseMerge(
pTab->bDescIdx, nDiff, pLeft, nLeft, &pRight, &nRight
);
sqlite3_free(pLeft);
p->doclist.aAll = pRight;
p->doclist.nAll = nRight;
}
if( iToken>p->iDoclistToken ) p->iDoclistToken = iToken;
return rc;
}
/*
** Load the doclist for phrase p into p->doclist.aAll/nAll. The loaded doclist
** does not take deferred tokens into account.
**
** SQLITE_OK is returned if no error occurs, otherwise an SQLite error code.
*/
static int fts3EvalPhraseLoad(
Fts3Cursor *pCsr, /* FTS Cursor handle */
Fts3Phrase *p /* Phrase object */
){
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
int iToken;
int rc = SQLITE_OK;
for(iToken=0; rc==SQLITE_OK && iToken<p->nToken; iToken++){
Fts3PhraseToken *pToken = &p->aToken[iToken];
assert( pToken->pDeferred==0 || pToken->pSegcsr==0 );
if( pToken->pSegcsr ){
int nThis = 0;
char *pThis = 0;
rc = fts3TermSelect(pTab, pToken, p->iColumn, &nThis, &pThis);
if( rc==SQLITE_OK ){
rc = fts3EvalPhraseMergeToken(pTab, p, iToken, pThis, nThis);
}
}
assert( pToken->pSegcsr==0 );
}
return rc;
}
#ifndef SQLITE_DISABLE_FTS4_DEFERRED
/*
** This function is called on each phrase after the position lists for
** any deferred tokens have been loaded into memory. It updates the phrases
** current position list to include only those positions that are really
** instances of the phrase (after considering deferred tokens). If this
** means that the phrase does not appear in the current row, doclist.pList
** and doclist.nList are both zeroed.
**
** SQLITE_OK is returned if no error occurs, otherwise an SQLite error code.
*/
static int fts3EvalDeferredPhrase(Fts3Cursor *pCsr, Fts3Phrase *pPhrase){
int iToken; /* Used to iterate through phrase tokens */
char *aPoslist = 0; /* Position list for deferred tokens */
int nPoslist = 0; /* Number of bytes in aPoslist */
int iPrev = -1; /* Token number of previous deferred token */
char *aFree = (pPhrase->doclist.bFreeList ? pPhrase->doclist.pList : 0);
for(iToken=0; iToken<pPhrase->nToken; iToken++){
Fts3PhraseToken *pToken = &pPhrase->aToken[iToken];
Fts3DeferredToken *pDeferred = pToken->pDeferred;
if( pDeferred ){
char *pList;
int nList;
int rc = sqlite3Fts3DeferredTokenList(pDeferred, &pList, &nList);
if( rc!=SQLITE_OK ) return rc;
if( pList==0 ){
sqlite3_free(aPoslist);
sqlite3_free(aFree);
pPhrase->doclist.pList = 0;
pPhrase->doclist.nList = 0;
return SQLITE_OK;
}else if( aPoslist==0 ){
aPoslist = pList;
nPoslist = nList;
}else{
char *aOut = pList;
char *p1 = aPoslist;
char *p2 = aOut;
assert( iPrev>=0 );
fts3PoslistPhraseMerge(&aOut, iToken-iPrev, 0, 1, &p1, &p2);
sqlite3_free(aPoslist);
aPoslist = pList;
nPoslist = (int)(aOut - aPoslist);
if( nPoslist==0 ){
sqlite3_free(aPoslist);
sqlite3_free(aFree);
pPhrase->doclist.pList = 0;
pPhrase->doclist.nList = 0;
return SQLITE_OK;
}
}
iPrev = iToken;
}
}
if( iPrev>=0 ){
int nMaxUndeferred = pPhrase->iDoclistToken;
if( nMaxUndeferred<0 ){
pPhrase->doclist.pList = aPoslist;
pPhrase->doclist.nList = nPoslist;
pPhrase->doclist.iDocid = pCsr->iPrevId;
pPhrase->doclist.bFreeList = 1;
}else{
int nDistance;
char *p1;
char *p2;
char *aOut;
if( nMaxUndeferred>iPrev ){
p1 = aPoslist;
p2 = pPhrase->doclist.pList;
nDistance = nMaxUndeferred - iPrev;
}else{
p1 = pPhrase->doclist.pList;
p2 = aPoslist;
nDistance = iPrev - nMaxUndeferred;
}
aOut = (char *)sqlite3Fts3MallocZero(nPoslist+FTS3_BUFFER_PADDING);
if( !aOut ){
sqlite3_free(aPoslist);
return SQLITE_NOMEM;
}
pPhrase->doclist.pList = aOut;
assert( p1 && p2 );
if( fts3PoslistPhraseMerge(&aOut, nDistance, 0, 1, &p1, &p2) ){
pPhrase->doclist.bFreeList = 1;
pPhrase->doclist.nList = (int)(aOut - pPhrase->doclist.pList);
}else{
sqlite3_free(aOut);
pPhrase->doclist.pList = 0;
pPhrase->doclist.nList = 0;
}
sqlite3_free(aPoslist);
}
}
if( pPhrase->doclist.pList!=aFree ) sqlite3_free(aFree);
return SQLITE_OK;
}
#endif /* SQLITE_DISABLE_FTS4_DEFERRED */
/*
** Maximum number of tokens a phrase may have to be considered for the
** incremental doclists strategy.
*/
#define MAX_INCR_PHRASE_TOKENS 4
/*
** This function is called for each Fts3Phrase in a full-text query
** expression to initialize the mechanism for returning rows. Once this
** function has been called successfully on an Fts3Phrase, it may be
** used with fts3EvalPhraseNext() to iterate through the matching docids.
**
** If parameter bOptOk is true, then the phrase may (or may not) use the
** incremental loading strategy. Otherwise, the entire doclist is loaded into
** memory within this call.
**
** SQLITE_OK is returned if no error occurs, otherwise an SQLite error code.
*/
static int fts3EvalPhraseStart(Fts3Cursor *pCsr, int bOptOk, Fts3Phrase *p){
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
int rc = SQLITE_OK; /* Error code */
int i;
/* Determine if doclists may be loaded from disk incrementally. This is
** possible if the bOptOk argument is true, the FTS doclists will be
** scanned in forward order, and the phrase consists of
** MAX_INCR_PHRASE_TOKENS or fewer tokens, none of which are are "^first"
** tokens or prefix tokens that cannot use a prefix-index. */
int bHaveIncr = 0;
int bIncrOk = (bOptOk
&& pCsr->bDesc==pTab->bDescIdx
&& p->nToken<=MAX_INCR_PHRASE_TOKENS && p->nToken>0
#if defined(SQLITE_DEBUG) || defined(SQLITE_TEST)
&& pTab->bNoIncrDoclist==0
#endif
);
for(i=0; bIncrOk==1 && i<p->nToken; i++){
Fts3PhraseToken *pToken = &p->aToken[i];
if( pToken->bFirst || (pToken->pSegcsr!=0 && !pToken->pSegcsr->bLookup) ){
bIncrOk = 0;
}
if( pToken->pSegcsr ) bHaveIncr = 1;
}
if( bIncrOk && bHaveIncr ){
/* Use the incremental approach. */
int iCol = (p->iColumn >= pTab->nColumn ? -1 : p->iColumn);
for(i=0; rc==SQLITE_OK && i<p->nToken; i++){
Fts3PhraseToken *pToken = &p->aToken[i];
Fts3MultiSegReader *pSegcsr = pToken->pSegcsr;
if( pSegcsr ){
rc = sqlite3Fts3MsrIncrStart(pTab, pSegcsr, iCol, pToken->z, pToken->n);
}
}
p->bIncr = 1;
}else{
/* Load the full doclist for the phrase into memory. */
rc = fts3EvalPhraseLoad(pCsr, p);
p->bIncr = 0;
}
assert( rc!=SQLITE_OK || p->nToken<1 || p->aToken[0].pSegcsr==0 || p->bIncr );
return rc;
}
/*
** This function is used to iterate backwards (from the end to start)
** through doclists. It is used by this module to iterate through phrase
** doclists in reverse and by the fts3_write.c module to iterate through
** pending-terms lists when writing to databases with "order=desc".
**
** The doclist may be sorted in ascending (parameter bDescIdx==0) or
** descending (parameter bDescIdx==1) order of docid. Regardless, this
** function iterates from the end of the doclist to the beginning.
*/
void sqlite3Fts3DoclistPrev(
int bDescIdx, /* True if the doclist is desc */
char *aDoclist, /* Pointer to entire doclist */
int nDoclist, /* Length of aDoclist in bytes */
char **ppIter, /* IN/OUT: Iterator pointer */
sqlite3_int64 *piDocid, /* IN/OUT: Docid pointer */
int *pnList, /* OUT: List length pointer */
u8 *pbEof /* OUT: End-of-file flag */
){
char *p = *ppIter;
assert( nDoclist>0 );
assert( *pbEof==0 );
assert_fts3_nc( p || *piDocid==0 );
assert( !p || (p>aDoclist && p<&aDoclist[nDoclist]) );
if( p==0 ){
sqlite3_int64 iDocid = 0;
char *pNext = 0;
char *pDocid = aDoclist;
char *pEnd = &aDoclist[nDoclist];
int iMul = 1;
while( pDocid<pEnd ){
sqlite3_int64 iDelta;
pDocid += sqlite3Fts3GetVarint(pDocid, &iDelta);
iDocid += (iMul * iDelta);
pNext = pDocid;
fts3PoslistCopy(0, &pDocid);
while( pDocid<pEnd && *pDocid==0 ) pDocid++;
iMul = (bDescIdx ? -1 : 1);
}
*pnList = (int)(pEnd - pNext);
*ppIter = pNext;
*piDocid = iDocid;
}else{
int iMul = (bDescIdx ? -1 : 1);
sqlite3_int64 iDelta;
fts3GetReverseVarint(&p, aDoclist, &iDelta);
*piDocid -= (iMul * iDelta);
if( p==aDoclist ){
*pbEof = 1;
}else{
char *pSave = p;
fts3ReversePoslist(aDoclist, &p);
*pnList = (int)(pSave - p);
}
*ppIter = p;
}
}
/*
** Iterate forwards through a doclist.
*/
void sqlite3Fts3DoclistNext(
int bDescIdx, /* True if the doclist is desc */
char *aDoclist, /* Pointer to entire doclist */
int nDoclist, /* Length of aDoclist in bytes */
char **ppIter, /* IN/OUT: Iterator pointer */
sqlite3_int64 *piDocid, /* IN/OUT: Docid pointer */
u8 *pbEof /* OUT: End-of-file flag */
){
char *p = *ppIter;
assert( nDoclist>0 );
assert( *pbEof==0 );
assert_fts3_nc( p || *piDocid==0 );
assert( !p || (p>=aDoclist && p<=&aDoclist[nDoclist]) );
if( p==0 ){
p = aDoclist;
p += sqlite3Fts3GetVarint(p, piDocid);
}else{
fts3PoslistCopy(0, &p);
while( p<&aDoclist[nDoclist] && *p==0 ) p++;
if( p>=&aDoclist[nDoclist] ){
*pbEof = 1;
}else{
sqlite3_int64 iVar;
p += sqlite3Fts3GetVarint(p, &iVar);
*piDocid += ((bDescIdx ? -1 : 1) * iVar);
}
}
*ppIter = p;
}
/*
** Advance the iterator pDL to the next entry in pDL->aAll/nAll. Set *pbEof
** to true if EOF is reached.
*/
static void fts3EvalDlPhraseNext(
Fts3Table *pTab,
Fts3Doclist *pDL,
u8 *pbEof
){
char *pIter; /* Used to iterate through aAll */
char *pEnd; /* 1 byte past end of aAll */
if( pDL->pNextDocid ){
pIter = pDL->pNextDocid;
assert( pDL->aAll!=0 || pIter==0 );
}else{
pIter = pDL->aAll;
}
if( pIter==0 || pIter>=(pEnd = pDL->aAll + pDL->nAll) ){
/* We have already reached the end of this doclist. EOF. */
*pbEof = 1;
}else{
sqlite3_int64 iDelta;
pIter += sqlite3Fts3GetVarint(pIter, &iDelta);
if( pTab->bDescIdx==0 || pDL->pNextDocid==0 ){
pDL->iDocid += iDelta;
}else{
pDL->iDocid -= iDelta;
}
pDL->pList = pIter;
fts3PoslistCopy(0, &pIter);
pDL->nList = (int)(pIter - pDL->pList);
/* pIter now points just past the 0x00 that terminates the position-
** list for document pDL->iDocid. However, if this position-list was
** edited in place by fts3EvalNearTrim(), then pIter may not actually
** point to the start of the next docid value. The following line deals
** with this case by advancing pIter past the zero-padding added by
** fts3EvalNearTrim(). */
while( pIter<pEnd && *pIter==0 ) pIter++;
pDL->pNextDocid = pIter;
assert( pIter>=&pDL->aAll[pDL->nAll] || *pIter );
*pbEof = 0;
}
}
/*
** Helper type used by fts3EvalIncrPhraseNext() and incrPhraseTokenNext().
*/
typedef struct TokenDoclist TokenDoclist;
struct TokenDoclist {
int bIgnore;
sqlite3_int64 iDocid;
char *pList;
int nList;
};
/*
** Token pToken is an incrementally loaded token that is part of a
** multi-token phrase. Advance it to the next matching document in the
** database and populate output variable *p with the details of the new
** entry. Or, if the iterator has reached EOF, set *pbEof to true.
**
** If an error occurs, return an SQLite error code. Otherwise, return
** SQLITE_OK.
*/
static int incrPhraseTokenNext(
Fts3Table *pTab, /* Virtual table handle */
Fts3Phrase *pPhrase, /* Phrase to advance token of */
int iToken, /* Specific token to advance */
TokenDoclist *p, /* OUT: Docid and doclist for new entry */
u8 *pbEof /* OUT: True if iterator is at EOF */
){
int rc = SQLITE_OK;
if( pPhrase->iDoclistToken==iToken ){
assert( p->bIgnore==0 );
assert( pPhrase->aToken[iToken].pSegcsr==0 );
fts3EvalDlPhraseNext(pTab, &pPhrase->doclist, pbEof);
p->pList = pPhrase->doclist.pList;
p->nList = pPhrase->doclist.nList;
p->iDocid = pPhrase->doclist.iDocid;
}else{
Fts3PhraseToken *pToken = &pPhrase->aToken[iToken];
assert( pToken->pDeferred==0 );
assert( pToken->pSegcsr || pPhrase->iDoclistToken>=0 );
if( pToken->pSegcsr ){
assert( p->bIgnore==0 );
rc = sqlite3Fts3MsrIncrNext(
pTab, pToken->pSegcsr, &p->iDocid, &p->pList, &p->nList
);
if( p->pList==0 ) *pbEof = 1;
}else{
p->bIgnore = 1;
}
}
return rc;
}
/*
** The phrase iterator passed as the second argument:
**
** * features at least one token that uses an incremental doclist, and
**
** * does not contain any deferred tokens.
**
** Advance it to the next matching documnent in the database and populate
** the Fts3Doclist.pList and nList fields.
**
** If there is no "next" entry and no error occurs, then *pbEof is set to
** 1 before returning. Otherwise, if no error occurs and the iterator is
** successfully advanced, *pbEof is set to 0.
**
** If an error occurs, return an SQLite error code. Otherwise, return
** SQLITE_OK.
*/
static int fts3EvalIncrPhraseNext(
Fts3Cursor *pCsr, /* FTS Cursor handle */
Fts3Phrase *p, /* Phrase object to advance to next docid */
u8 *pbEof /* OUT: Set to 1 if EOF */
){
int rc = SQLITE_OK;
Fts3Doclist *pDL = &p->doclist;
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
u8 bEof = 0;
/* This is only called if it is guaranteed that the phrase has at least
** one incremental token. In which case the bIncr flag is set. */
assert( p->bIncr==1 );
if( p->nToken==1 ){
rc = sqlite3Fts3MsrIncrNext(pTab, p->aToken[0].pSegcsr,
&pDL->iDocid, &pDL->pList, &pDL->nList
);
if( pDL->pList==0 ) bEof = 1;
}else{
int bDescDoclist = pCsr->bDesc;
struct TokenDoclist a[MAX_INCR_PHRASE_TOKENS];
memset(a, 0, sizeof(a));
assert( p->nToken<=MAX_INCR_PHRASE_TOKENS );
assert( p->iDoclistToken<MAX_INCR_PHRASE_TOKENS );
while( bEof==0 ){
int bMaxSet = 0;
sqlite3_int64 iMax = 0; /* Largest docid for all iterators */
int i; /* Used to iterate through tokens */
/* Advance the iterator for each token in the phrase once. */
for(i=0; rc==SQLITE_OK && i<p->nToken && bEof==0; i++){
rc = incrPhraseTokenNext(pTab, p, i, &a[i], &bEof);
if( a[i].bIgnore==0 && (bMaxSet==0 || DOCID_CMP(iMax, a[i].iDocid)<0) ){
iMax = a[i].iDocid;
bMaxSet = 1;
}
}
assert( rc!=SQLITE_OK || (p->nToken>=1 && a[p->nToken-1].bIgnore==0) );
assert( rc!=SQLITE_OK || bMaxSet );
/* Keep advancing iterators until they all point to the same document */
for(i=0; i<p->nToken; i++){
while( rc==SQLITE_OK && bEof==0
&& a[i].bIgnore==0 && DOCID_CMP(a[i].iDocid, iMax)<0
){
rc = incrPhraseTokenNext(pTab, p, i, &a[i], &bEof);
if( DOCID_CMP(a[i].iDocid, iMax)>0 ){
iMax = a[i].iDocid;
i = 0;
}
}
}
/* Check if the current entries really are a phrase match */
if( bEof==0 ){
int nList = 0;
int nByte = a[p->nToken-1].nList;
char *aDoclist = sqlite3_malloc64((i64)nByte+FTS3_BUFFER_PADDING);
if( !aDoclist ) return SQLITE_NOMEM;
memcpy(aDoclist, a[p->nToken-1].pList, nByte+1);
memset(&aDoclist[nByte], 0, FTS3_BUFFER_PADDING);
for(i=0; i<(p->nToken-1); i++){
if( a[i].bIgnore==0 ){
char *pL = a[i].pList;
char *pR = aDoclist;
char *pOut = aDoclist;
int nDist = p->nToken-1-i;
int res = fts3PoslistPhraseMerge(&pOut, nDist, 0, 1, &pL, &pR);
if( res==0 ) break;
nList = (int)(pOut - aDoclist);
}
}
if( i==(p->nToken-1) ){
pDL->iDocid = iMax;
pDL->pList = aDoclist;
pDL->nList = nList;
pDL->bFreeList = 1;
break;
}
sqlite3_free(aDoclist);
}
}
}
*pbEof = bEof;
return rc;
}
/*
** Attempt to move the phrase iterator to point to the next matching docid.
** If an error occurs, return an SQLite error code. Otherwise, return
** SQLITE_OK.
**
** If there is no "next" entry and no error occurs, then *pbEof is set to
** 1 before returning. Otherwise, if no error occurs and the iterator is
** successfully advanced, *pbEof is set to 0.
*/
static int fts3EvalPhraseNext(
Fts3Cursor *pCsr, /* FTS Cursor handle */
Fts3Phrase *p, /* Phrase object to advance to next docid */
u8 *pbEof /* OUT: Set to 1 if EOF */
){
int rc = SQLITE_OK;
Fts3Doclist *pDL = &p->doclist;
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
if( p->bIncr ){
rc = fts3EvalIncrPhraseNext(pCsr, p, pbEof);
}else if( pCsr->bDesc!=pTab->bDescIdx && pDL->nAll ){
sqlite3Fts3DoclistPrev(pTab->bDescIdx, pDL->aAll, pDL->nAll,
&pDL->pNextDocid, &pDL->iDocid, &pDL->nList, pbEof
);
pDL->pList = pDL->pNextDocid;
}else{
fts3EvalDlPhraseNext(pTab, pDL, pbEof);
}
return rc;
}
/*
**
** If *pRc is not SQLITE_OK when this function is called, it is a no-op.
** Otherwise, fts3EvalPhraseStart() is called on all phrases within the
** expression. Also the Fts3Expr.bDeferred variable is set to true for any
** expressions for which all descendent tokens are deferred.
**
** If parameter bOptOk is zero, then it is guaranteed that the
** Fts3Phrase.doclist.aAll/nAll variables contain the entire doclist for
** each phrase in the expression (subject to deferred token processing).
** Or, if bOptOk is non-zero, then one or more tokens within the expression
** may be loaded incrementally, meaning doclist.aAll/nAll is not available.
**
** If an error occurs within this function, *pRc is set to an SQLite error
** code before returning.
*/
static void fts3EvalStartReaders(
Fts3Cursor *pCsr, /* FTS Cursor handle */
Fts3Expr *pExpr, /* Expression to initialize phrases in */
int *pRc /* IN/OUT: Error code */
){
if( pExpr && SQLITE_OK==*pRc ){
if( pExpr->eType==FTSQUERY_PHRASE ){
int nToken = pExpr->pPhrase->nToken;
if( nToken ){
int i;
for(i=0; i<nToken; i++){
if( pExpr->pPhrase->aToken[i].pDeferred==0 ) break;
}
pExpr->bDeferred = (i==nToken);
}
*pRc = fts3EvalPhraseStart(pCsr, 1, pExpr->pPhrase);
}else{
fts3EvalStartReaders(pCsr, pExpr->pLeft, pRc);
fts3EvalStartReaders(pCsr, pExpr->pRight, pRc);
pExpr->bDeferred = (pExpr->pLeft->bDeferred && pExpr->pRight->bDeferred);
}
}
}
/*
** An array of the following structures is assembled as part of the process
** of selecting tokens to defer before the query starts executing (as part
** of the xFilter() method). There is one element in the array for each
** token in the FTS expression.
**
** Tokens are divided into AND/NEAR clusters. All tokens in a cluster belong
** to phrases that are connected only by AND and NEAR operators (not OR or
** NOT). When determining tokens to defer, each AND/NEAR cluster is considered
** separately. The root of a tokens AND/NEAR cluster is stored in
** Fts3TokenAndCost.pRoot.
*/
typedef struct Fts3TokenAndCost Fts3TokenAndCost;
struct Fts3TokenAndCost {
Fts3Phrase *pPhrase; /* The phrase the token belongs to */
int iToken; /* Position of token in phrase */
Fts3PhraseToken *pToken; /* The token itself */
Fts3Expr *pRoot; /* Root of NEAR/AND cluster */
int nOvfl; /* Number of overflow pages to load doclist */
int iCol; /* The column the token must match */
};
/*
** This function is used to populate an allocated Fts3TokenAndCost array.
**
** If *pRc is not SQLITE_OK when this function is called, it is a no-op.
** Otherwise, if an error occurs during execution, *pRc is set to an
** SQLite error code.
*/
static void fts3EvalTokenCosts(
Fts3Cursor *pCsr, /* FTS Cursor handle */
Fts3Expr *pRoot, /* Root of current AND/NEAR cluster */
Fts3Expr *pExpr, /* Expression to consider */
Fts3TokenAndCost **ppTC, /* Write new entries to *(*ppTC)++ */
Fts3Expr ***ppOr, /* Write new OR root to *(*ppOr)++ */
int *pRc /* IN/OUT: Error code */
){
if( *pRc==SQLITE_OK ){
if( pExpr->eType==FTSQUERY_PHRASE ){
Fts3Phrase *pPhrase = pExpr->pPhrase;
int i;
for(i=0; *pRc==SQLITE_OK && i<pPhrase->nToken; i++){
Fts3TokenAndCost *pTC = (*ppTC)++;
pTC->pPhrase = pPhrase;
pTC->iToken = i;
pTC->pRoot = pRoot;
pTC->pToken = &pPhrase->aToken[i];
pTC->iCol = pPhrase->iColumn;
*pRc = sqlite3Fts3MsrOvfl(pCsr, pTC->pToken->pSegcsr, &pTC->nOvfl);
}
}else if( pExpr->eType!=FTSQUERY_NOT ){
assert( pExpr->eType==FTSQUERY_OR
|| pExpr->eType==FTSQUERY_AND
|| pExpr->eType==FTSQUERY_NEAR
);
assert( pExpr->pLeft && pExpr->pRight );
if( pExpr->eType==FTSQUERY_OR ){
pRoot = pExpr->pLeft;
**ppOr = pRoot;
(*ppOr)++;
}
fts3EvalTokenCosts(pCsr, pRoot, pExpr->pLeft, ppTC, ppOr, pRc);
if( pExpr->eType==FTSQUERY_OR ){
pRoot = pExpr->pRight;
**ppOr = pRoot;
(*ppOr)++;
}
fts3EvalTokenCosts(pCsr, pRoot, pExpr->pRight, ppTC, ppOr, pRc);
}
}
}
/*
** Determine the average document (row) size in pages. If successful,
** write this value to *pnPage and return SQLITE_OK. Otherwise, return
** an SQLite error code.
**
** The average document size in pages is calculated by first calculating
** determining the average size in bytes, B. If B is less than the amount
** of data that will fit on a single leaf page of an intkey table in
** this database, then the average docsize is 1. Otherwise, it is 1 plus
** the number of overflow pages consumed by a record B bytes in size.
*/
static int fts3EvalAverageDocsize(Fts3Cursor *pCsr, int *pnPage){
int rc = SQLITE_OK;
if( pCsr->nRowAvg==0 ){
/* The average document size, which is required to calculate the cost
** of each doclist, has not yet been determined. Read the required
** data from the %_stat table to calculate it.
**
** Entry 0 of the %_stat table is a blob containing (nCol+1) FTS3
** varints, where nCol is the number of columns in the FTS3 table.
** The first varint is the number of documents currently stored in
** the table. The following nCol varints contain the total amount of
** data stored in all rows of each column of the table, from left
** to right.
*/
Fts3Table *p = (Fts3Table*)pCsr->base.pVtab;
sqlite3_stmt *pStmt;
sqlite3_int64 nDoc = 0;
sqlite3_int64 nByte = 0;
const char *pEnd;
const char *a;
rc = sqlite3Fts3SelectDoctotal(p, &pStmt);
if( rc!=SQLITE_OK ) return rc;
a = sqlite3_column_blob(pStmt, 0);
testcase( a==0 ); /* If %_stat.value set to X'' */
if( a ){
pEnd = &a[sqlite3_column_bytes(pStmt, 0)];
a += sqlite3Fts3GetVarintBounded(a, pEnd, &nDoc);
while( a<pEnd ){
a += sqlite3Fts3GetVarintBounded(a, pEnd, &nByte);
}
}
if( nDoc==0 || nByte==0 ){
sqlite3_reset(pStmt);
return FTS_CORRUPT_VTAB;
}
pCsr->nDoc = nDoc;
pCsr->nRowAvg = (int)(((nByte / nDoc) + p->nPgsz) / p->nPgsz);
assert( pCsr->nRowAvg>0 );
rc = sqlite3_reset(pStmt);
}
*pnPage = pCsr->nRowAvg;
return rc;
}
/*
** This function is called to select the tokens (if any) that will be
** deferred. The array aTC[] has already been populated when this is
** called.
**
** This function is called once for each AND/NEAR cluster in the
** expression. Each invocation determines which tokens to defer within
** the cluster with root node pRoot. See comments above the definition
** of struct Fts3TokenAndCost for more details.
**
** If no error occurs, SQLITE_OK is returned and sqlite3Fts3DeferToken()
** called on each token to defer. Otherwise, an SQLite error code is
** returned.
*/
static int fts3EvalSelectDeferred(
Fts3Cursor *pCsr, /* FTS Cursor handle */
Fts3Expr *pRoot, /* Consider tokens with this root node */
Fts3TokenAndCost *aTC, /* Array of expression tokens and costs */
int nTC /* Number of entries in aTC[] */
){
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
int nDocSize = 0; /* Number of pages per doc loaded */
int rc = SQLITE_OK; /* Return code */
int ii; /* Iterator variable for various purposes */
int nOvfl = 0; /* Total overflow pages used by doclists */
int nToken = 0; /* Total number of tokens in cluster */
int nMinEst = 0; /* The minimum count for any phrase so far. */
int nLoad4 = 1; /* (Phrases that will be loaded)^4. */
/* Tokens are never deferred for FTS tables created using the content=xxx
** option. The reason being that it is not guaranteed that the content
** table actually contains the same data as the index. To prevent this from
** causing any problems, the deferred token optimization is completely
** disabled for content=xxx tables. */
if( pTab->zContentTbl ){
return SQLITE_OK;
}
/* Count the tokens in this AND/NEAR cluster. If none of the doclists
** associated with the tokens spill onto overflow pages, or if there is
** only 1 token, exit early. No tokens to defer in this case. */
for(ii=0; ii<nTC; ii++){
if( aTC[ii].pRoot==pRoot ){
nOvfl += aTC[ii].nOvfl;
nToken++;
}
}
if( nOvfl==0 || nToken<2 ) return SQLITE_OK;
/* Obtain the average docsize (in pages). */
rc = fts3EvalAverageDocsize(pCsr, &nDocSize);
assert( rc!=SQLITE_OK || nDocSize>0 );
/* Iterate through all tokens in this AND/NEAR cluster, in ascending order
** of the number of overflow pages that will be loaded by the pager layer
** to retrieve the entire doclist for the token from the full-text index.
** Load the doclists for tokens that are either:
**
** a. The cheapest token in the entire query (i.e. the one visited by the
** first iteration of this loop), or
**
** b. Part of a multi-token phrase.
**
** After each token doclist is loaded, merge it with the others from the
** same phrase and count the number of documents that the merged doclist
** contains. Set variable "nMinEst" to the smallest number of documents in
** any phrase doclist for which 1 or more token doclists have been loaded.
** Let nOther be the number of other phrases for which it is certain that
** one or more tokens will not be deferred.
**
** Then, for each token, defer it if loading the doclist would result in
** loading N or more overflow pages into memory, where N is computed as:
**
** (nMinEst + 4^nOther - 1) / (4^nOther)
*/
for(ii=0; ii<nToken && rc==SQLITE_OK; ii++){
int iTC; /* Used to iterate through aTC[] array. */
Fts3TokenAndCost *pTC = 0; /* Set to cheapest remaining token. */
/* Set pTC to point to the cheapest remaining token. */
for(iTC=0; iTC<nTC; iTC++){
if( aTC[iTC].pToken && aTC[iTC].pRoot==pRoot
&& (!pTC || aTC[iTC].nOvfl<pTC->nOvfl)
){
pTC = &aTC[iTC];
}
}
assert( pTC );
if( ii && pTC->nOvfl>=((nMinEst+(nLoad4/4)-1)/(nLoad4/4))*nDocSize ){
/* The number of overflow pages to load for this (and therefore all
** subsequent) tokens is greater than the estimated number of pages
** that will be loaded if all subsequent tokens are deferred.
*/
Fts3PhraseToken *pToken = pTC->pToken;
rc = sqlite3Fts3DeferToken(pCsr, pToken, pTC->iCol);
fts3SegReaderCursorFree(pToken->pSegcsr);
pToken->pSegcsr = 0;
}else{
/* Set nLoad4 to the value of (4^nOther) for the next iteration of the
** for-loop. Except, limit the value to 2^24 to prevent it from
** overflowing the 32-bit integer it is stored in. */
if( ii<12 ) nLoad4 = nLoad4*4;
if( ii==0 || (pTC->pPhrase->nToken>1 && ii!=nToken-1) ){
/* Either this is the cheapest token in the entire query, or it is
** part of a multi-token phrase. Either way, the entire doclist will
** (eventually) be loaded into memory. It may as well be now. */
Fts3PhraseToken *pToken = pTC->pToken;
int nList = 0;
char *pList = 0;
rc = fts3TermSelect(pTab, pToken, pTC->iCol, &nList, &pList);
assert( rc==SQLITE_OK || pList==0 );
if( rc==SQLITE_OK ){
rc = fts3EvalPhraseMergeToken(
pTab, pTC->pPhrase, pTC->iToken,pList,nList
);
}
if( rc==SQLITE_OK ){
int nCount;
nCount = fts3DoclistCountDocids(
pTC->pPhrase->doclist.aAll, pTC->pPhrase->doclist.nAll
);
if( ii==0 || nCount<nMinEst ) nMinEst = nCount;
}
}
}
pTC->pToken = 0;
}
return rc;
}
/*
** This function is called from within the xFilter method. It initializes
** the full-text query currently stored in pCsr->pExpr. To iterate through
** the results of a query, the caller does:
**
** fts3EvalStart(pCsr);
** while( 1 ){
** fts3EvalNext(pCsr);
** if( pCsr->bEof ) break;
** ... return row pCsr->iPrevId to the caller ...
** }
*/
static int fts3EvalStart(Fts3Cursor *pCsr){
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
int rc = SQLITE_OK;
int nToken = 0;
int nOr = 0;
/* Allocate a MultiSegReader for each token in the expression. */
fts3EvalAllocateReaders(pCsr, pCsr->pExpr, &nToken, &nOr, &rc);
/* Determine which, if any, tokens in the expression should be deferred. */
#ifndef SQLITE_DISABLE_FTS4_DEFERRED
if( rc==SQLITE_OK && nToken>1 && pTab->bFts4 ){
Fts3TokenAndCost *aTC;
aTC = (Fts3TokenAndCost *)sqlite3_malloc64(
sizeof(Fts3TokenAndCost) * nToken
+ sizeof(Fts3Expr *) * nOr * 2
);
if( !aTC ){
rc = SQLITE_NOMEM;
}else{
Fts3Expr **apOr = (Fts3Expr **)&aTC[nToken];
int ii;
Fts3TokenAndCost *pTC = aTC;
Fts3Expr **ppOr = apOr;
fts3EvalTokenCosts(pCsr, 0, pCsr->pExpr, &pTC, &ppOr, &rc);
nToken = (int)(pTC-aTC);
nOr = (int)(ppOr-apOr);
if( rc==SQLITE_OK ){
rc = fts3EvalSelectDeferred(pCsr, 0, aTC, nToken);
for(ii=0; rc==SQLITE_OK && ii<nOr; ii++){
rc = fts3EvalSelectDeferred(pCsr, apOr[ii], aTC, nToken);
}
}
sqlite3_free(aTC);
}
}
#endif
fts3EvalStartReaders(pCsr, pCsr->pExpr, &rc);
return rc;
}
/*
** Invalidate the current position list for phrase pPhrase.
*/
static void fts3EvalInvalidatePoslist(Fts3Phrase *pPhrase){
if( pPhrase->doclist.bFreeList ){
sqlite3_free(pPhrase->doclist.pList);
}
pPhrase->doclist.pList = 0;
pPhrase->doclist.nList = 0;
pPhrase->doclist.bFreeList = 0;
}
/*
** This function is called to edit the position list associated with
** the phrase object passed as the fifth argument according to a NEAR
** condition. For example:
**
** abc NEAR/5 "def ghi"
**
** Parameter nNear is passed the NEAR distance of the expression (5 in
** the example above). When this function is called, *paPoslist points to
** the position list, and *pnToken is the number of phrase tokens in the
** phrase on the other side of the NEAR operator to pPhrase. For example,
** if pPhrase refers to the "def ghi" phrase, then *paPoslist points to
** the position list associated with phrase "abc".
**
** All positions in the pPhrase position list that are not sufficiently
** close to a position in the *paPoslist position list are removed. If this
** leaves 0 positions, zero is returned. Otherwise, non-zero.
**
** Before returning, *paPoslist is set to point to the position lsit
** associated with pPhrase. And *pnToken is set to the number of tokens in
** pPhrase.
*/
static int fts3EvalNearTrim(
int nNear, /* NEAR distance. As in "NEAR/nNear". */
char *aTmp, /* Temporary space to use */
char **paPoslist, /* IN/OUT: Position list */
int *pnToken, /* IN/OUT: Tokens in phrase of *paPoslist */
Fts3Phrase *pPhrase /* The phrase object to trim the doclist of */
){
int nParam1 = nNear + pPhrase->nToken;
int nParam2 = nNear + *pnToken;
int nNew;
char *p2;
char *pOut;
int res;
assert( pPhrase->doclist.pList );
p2 = pOut = pPhrase->doclist.pList;
res = fts3PoslistNearMerge(
&pOut, aTmp, nParam1, nParam2, paPoslist, &p2
);
if( res ){
nNew = (int)(pOut - pPhrase->doclist.pList) - 1;
assert_fts3_nc( nNew<=pPhrase->doclist.nList && nNew>0 );
if( nNew>=0 && nNew<=pPhrase->doclist.nList ){
assert( pPhrase->doclist.pList[nNew]=='\0' );
memset(&pPhrase->doclist.pList[nNew], 0, pPhrase->doclist.nList - nNew);
pPhrase->doclist.nList = nNew;
}
*paPoslist = pPhrase->doclist.pList;
*pnToken = pPhrase->nToken;
}
return res;
}
/*
** This function is a no-op if *pRc is other than SQLITE_OK when it is called.
** Otherwise, it advances the expression passed as the second argument to
** point to the next matching row in the database. Expressions iterate through
** matching rows in docid order. Ascending order if Fts3Cursor.bDesc is zero,
** or descending if it is non-zero.
**
** If an error occurs, *pRc is set to an SQLite error code. Otherwise, if
** successful, the following variables in pExpr are set:
**
** Fts3Expr.bEof (non-zero if EOF - there is no next row)
** Fts3Expr.iDocid (valid if bEof==0. The docid of the next row)
**
** If the expression is of type FTSQUERY_PHRASE, and the expression is not
** at EOF, then the following variables are populated with the position list
** for the phrase for the visited row:
**
** FTs3Expr.pPhrase->doclist.nList (length of pList in bytes)
** FTs3Expr.pPhrase->doclist.pList (pointer to position list)
**
** It says above that this function advances the expression to the next
** matching row. This is usually true, but there are the following exceptions:
**
** 1. Deferred tokens are not taken into account. If a phrase consists
** entirely of deferred tokens, it is assumed to match every row in
** the db. In this case the position-list is not populated at all.
**
** Or, if a phrase contains one or more deferred tokens and one or
** more non-deferred tokens, then the expression is advanced to the
** next possible match, considering only non-deferred tokens. In other
** words, if the phrase is "A B C", and "B" is deferred, the expression
** is advanced to the next row that contains an instance of "A * C",
** where "*" may match any single token. The position list in this case
** is populated as for "A * C" before returning.
**
** 2. NEAR is treated as AND. If the expression is "x NEAR y", it is
** advanced to point to the next row that matches "x AND y".
**
** See sqlite3Fts3EvalTestDeferred() for details on testing if a row is
** really a match, taking into account deferred tokens and NEAR operators.
*/
static void fts3EvalNextRow(
Fts3Cursor *pCsr, /* FTS Cursor handle */
Fts3Expr *pExpr, /* Expr. to advance to next matching row */
int *pRc /* IN/OUT: Error code */
){
if( *pRc==SQLITE_OK && pExpr->bEof==0 ){
int bDescDoclist = pCsr->bDesc; /* Used by DOCID_CMP() macro */
pExpr->bStart = 1;
switch( pExpr->eType ){
case FTSQUERY_NEAR:
case FTSQUERY_AND: {
Fts3Expr *pLeft = pExpr->pLeft;
Fts3Expr *pRight = pExpr->pRight;
assert( !pLeft->bDeferred || !pRight->bDeferred );
if( pLeft->bDeferred ){
/* LHS is entirely deferred. So we assume it matches every row.
** Advance the RHS iterator to find the next row visited. */
fts3EvalNextRow(pCsr, pRight, pRc);
pExpr->iDocid = pRight->iDocid;
pExpr->bEof = pRight->bEof;
}else if( pRight->bDeferred ){
/* RHS is entirely deferred. So we assume it matches every row.
** Advance the LHS iterator to find the next row visited. */
fts3EvalNextRow(pCsr, pLeft, pRc);
pExpr->iDocid = pLeft->iDocid;
pExpr->bEof = pLeft->bEof;
}else{
/* Neither the RHS or LHS are deferred. */
fts3EvalNextRow(pCsr, pLeft, pRc);
fts3EvalNextRow(pCsr, pRight, pRc);
while( !pLeft->bEof && !pRight->bEof && *pRc==SQLITE_OK ){
sqlite3_int64 iDiff = DOCID_CMP(pLeft->iDocid, pRight->iDocid);
if( iDiff==0 ) break;
if( iDiff<0 ){
fts3EvalNextRow(pCsr, pLeft, pRc);
}else{
fts3EvalNextRow(pCsr, pRight, pRc);
}
}
pExpr->iDocid = pLeft->iDocid;
pExpr->bEof = (pLeft->bEof || pRight->bEof);
if( pExpr->eType==FTSQUERY_NEAR && pExpr->bEof ){
assert( pRight->eType==FTSQUERY_PHRASE );
if( pRight->pPhrase->doclist.aAll ){
Fts3Doclist *pDl = &pRight->pPhrase->doclist;
while( *pRc==SQLITE_OK && pRight->bEof==0 ){
memset(pDl->pList, 0, pDl->nList);
fts3EvalNextRow(pCsr, pRight, pRc);
}
}
if( pLeft->pPhrase && pLeft->pPhrase->doclist.aAll ){
Fts3Doclist *pDl = &pLeft->pPhrase->doclist;
while( *pRc==SQLITE_OK && pLeft->bEof==0 ){
memset(pDl->pList, 0, pDl->nList);
fts3EvalNextRow(pCsr, pLeft, pRc);
}
}
pRight->bEof = pLeft->bEof = 1;
}
}
break;
}
case FTSQUERY_OR: {
Fts3Expr *pLeft = pExpr->pLeft;
Fts3Expr *pRight = pExpr->pRight;
sqlite3_int64 iCmp = DOCID_CMP(pLeft->iDocid, pRight->iDocid);
assert_fts3_nc( pLeft->bStart || pLeft->iDocid==pRight->iDocid );
assert_fts3_nc( pRight->bStart || pLeft->iDocid==pRight->iDocid );
if( pRight->bEof || (pLeft->bEof==0 && iCmp<0) ){
fts3EvalNextRow(pCsr, pLeft, pRc);
}else if( pLeft->bEof || iCmp>0 ){
fts3EvalNextRow(pCsr, pRight, pRc);
}else{
fts3EvalNextRow(pCsr, pLeft, pRc);
fts3EvalNextRow(pCsr, pRight, pRc);
}
pExpr->bEof = (pLeft->bEof && pRight->bEof);
iCmp = DOCID_CMP(pLeft->iDocid, pRight->iDocid);
if( pRight->bEof || (pLeft->bEof==0 && iCmp<0) ){
pExpr->iDocid = pLeft->iDocid;
}else{
pExpr->iDocid = pRight->iDocid;
}
break;
}
case FTSQUERY_NOT: {
Fts3Expr *pLeft = pExpr->pLeft;
Fts3Expr *pRight = pExpr->pRight;
if( pRight->bStart==0 ){
fts3EvalNextRow(pCsr, pRight, pRc);
assert( *pRc!=SQLITE_OK || pRight->bStart );
}
fts3EvalNextRow(pCsr, pLeft, pRc);
if( pLeft->bEof==0 ){
while( !*pRc
&& !pRight->bEof
&& DOCID_CMP(pLeft->iDocid, pRight->iDocid)>0
){
fts3EvalNextRow(pCsr, pRight, pRc);
}
}
pExpr->iDocid = pLeft->iDocid;
pExpr->bEof = pLeft->bEof;
break;
}
default: {
Fts3Phrase *pPhrase = pExpr->pPhrase;
fts3EvalInvalidatePoslist(pPhrase);
*pRc = fts3EvalPhraseNext(pCsr, pPhrase, &pExpr->bEof);
pExpr->iDocid = pPhrase->doclist.iDocid;
break;
}
}
}
}
/*
** If *pRc is not SQLITE_OK, or if pExpr is not the root node of a NEAR
** cluster, then this function returns 1 immediately.
**
** Otherwise, it checks if the current row really does match the NEAR
** expression, using the data currently stored in the position lists
** (Fts3Expr->pPhrase.doclist.pList/nList) for each phrase in the expression.
**
** If the current row is a match, the position list associated with each
** phrase in the NEAR expression is edited in place to contain only those
** phrase instances sufficiently close to their peers to satisfy all NEAR
** constraints. In this case it returns 1. If the NEAR expression does not
** match the current row, 0 is returned. The position lists may or may not
** be edited if 0 is returned.
*/
static int fts3EvalNearTest(Fts3Expr *pExpr, int *pRc){
int res = 1;
/* The following block runs if pExpr is the root of a NEAR query.
** For example, the query:
**
** "w" NEAR "x" NEAR "y" NEAR "z"
**
** which is represented in tree form as:
**
** |
** +--NEAR--+ <-- root of NEAR query
** | |
** +--NEAR--+ "z"
** | |
** +--NEAR--+ "y"
** | |
** "w" "x"
**
** The right-hand child of a NEAR node is always a phrase. The
** left-hand child may be either a phrase or a NEAR node. There are
** no exceptions to this - it's the way the parser in fts3_expr.c works.
*/
if( *pRc==SQLITE_OK
&& pExpr->eType==FTSQUERY_NEAR
&& (pExpr->pParent==0 || pExpr->pParent->eType!=FTSQUERY_NEAR)
){
Fts3Expr *p;
sqlite3_int64 nTmp = 0; /* Bytes of temp space */
char *aTmp; /* Temp space for PoslistNearMerge() */
/* Allocate temporary working space. */
for(p=pExpr; p->pLeft; p=p->pLeft){
assert( p->pRight->pPhrase->doclist.nList>0 );
nTmp += p->pRight->pPhrase->doclist.nList;
}
nTmp += p->pPhrase->doclist.nList;
aTmp = sqlite3_malloc64(nTmp*2);
if( !aTmp ){
*pRc = SQLITE_NOMEM;
res = 0;
}else{
char *aPoslist = p->pPhrase->doclist.pList;
int nToken = p->pPhrase->nToken;
for(p=p->pParent;res && p && p->eType==FTSQUERY_NEAR; p=p->pParent){
Fts3Phrase *pPhrase = p->pRight->pPhrase;
int nNear = p->nNear;
res = fts3EvalNearTrim(nNear, aTmp, &aPoslist, &nToken, pPhrase);
}
aPoslist = pExpr->pRight->pPhrase->doclist.pList;
nToken = pExpr->pRight->pPhrase->nToken;
for(p=pExpr->pLeft; p && res; p=p->pLeft){
int nNear;
Fts3Phrase *pPhrase;
assert( p->pParent && p->pParent->pLeft==p );
nNear = p->pParent->nNear;
pPhrase = (
p->eType==FTSQUERY_NEAR ? p->pRight->pPhrase : p->pPhrase
);
res = fts3EvalNearTrim(nNear, aTmp, &aPoslist, &nToken, pPhrase);
}
}
sqlite3_free(aTmp);
}
return res;
}
/*
** This function is a helper function for sqlite3Fts3EvalTestDeferred().
** Assuming no error occurs or has occurred, It returns non-zero if the
** expression passed as the second argument matches the row that pCsr
** currently points to, or zero if it does not.
**
** If *pRc is not SQLITE_OK when this function is called, it is a no-op.
** If an error occurs during execution of this function, *pRc is set to
** the appropriate SQLite error code. In this case the returned value is
** undefined.
*/
static int fts3EvalTestExpr(
Fts3Cursor *pCsr, /* FTS cursor handle */
Fts3Expr *pExpr, /* Expr to test. May or may not be root. */
int *pRc /* IN/OUT: Error code */
){
int bHit = 1; /* Return value */
if( *pRc==SQLITE_OK ){
switch( pExpr->eType ){
case FTSQUERY_NEAR:
case FTSQUERY_AND:
bHit = (
fts3EvalTestExpr(pCsr, pExpr->pLeft, pRc)
&& fts3EvalTestExpr(pCsr, pExpr->pRight, pRc)
&& fts3EvalNearTest(pExpr, pRc)
);
/* If the NEAR expression does not match any rows, zero the doclist for
** all phrases involved in the NEAR. This is because the snippet(),
** offsets() and matchinfo() functions are not supposed to recognize
** any instances of phrases that are part of unmatched NEAR queries.
** For example if this expression:
**
** ... MATCH 'a OR (b NEAR c)'
**
** is matched against a row containing:
**
** 'a b d e'
**
** then any snippet() should ony highlight the "a" term, not the "b"
** (as "b" is part of a non-matching NEAR clause).
*/
if( bHit==0
&& pExpr->eType==FTSQUERY_NEAR
&& (pExpr->pParent==0 || pExpr->pParent->eType!=FTSQUERY_NEAR)
){
Fts3Expr *p;
for(p=pExpr; p->pPhrase==0; p=p->pLeft){
if( p->pRight->iDocid==pCsr->iPrevId ){
fts3EvalInvalidatePoslist(p->pRight->pPhrase);
}
}
if( p->iDocid==pCsr->iPrevId ){
fts3EvalInvalidatePoslist(p->pPhrase);
}
}
break;
case FTSQUERY_OR: {
int bHit1 = fts3EvalTestExpr(pCsr, pExpr->pLeft, pRc);
int bHit2 = fts3EvalTestExpr(pCsr, pExpr->pRight, pRc);
bHit = bHit1 || bHit2;
break;
}
case FTSQUERY_NOT:
bHit = (
fts3EvalTestExpr(pCsr, pExpr->pLeft, pRc)
&& !fts3EvalTestExpr(pCsr, pExpr->pRight, pRc)
);
break;
default: {
#ifndef SQLITE_DISABLE_FTS4_DEFERRED
if( pCsr->pDeferred && (pExpr->bDeferred || (
pExpr->iDocid==pCsr->iPrevId && pExpr->pPhrase->doclist.pList
))){
Fts3Phrase *pPhrase = pExpr->pPhrase;
if( pExpr->bDeferred ){
fts3EvalInvalidatePoslist(pPhrase);
}
*pRc = fts3EvalDeferredPhrase(pCsr, pPhrase);
bHit = (pPhrase->doclist.pList!=0);
pExpr->iDocid = pCsr->iPrevId;
}else
#endif
{
bHit = (
pExpr->bEof==0 && pExpr->iDocid==pCsr->iPrevId
&& pExpr->pPhrase->doclist.nList>0
);
}
break;
}
}
}
return bHit;
}
/*
** This function is called as the second part of each xNext operation when
** iterating through the results of a full-text query. At this point the
** cursor points to a row that matches the query expression, with the
** following caveats:
**
** * Up until this point, "NEAR" operators in the expression have been
** treated as "AND".
**
** * Deferred tokens have not yet been considered.
**
** If *pRc is not SQLITE_OK when this function is called, it immediately
** returns 0. Otherwise, it tests whether or not after considering NEAR
** operators and deferred tokens the current row is still a match for the
** expression. It returns 1 if both of the following are true:
**
** 1. *pRc is SQLITE_OK when this function returns, and
**
** 2. After scanning the current FTS table row for the deferred tokens,
** it is determined that the row does *not* match the query.
**
** Or, if no error occurs and it seems the current row does match the FTS
** query, return 0.
*/
int sqlite3Fts3EvalTestDeferred(Fts3Cursor *pCsr, int *pRc){
int rc = *pRc;
int bMiss = 0;
if( rc==SQLITE_OK ){
/* If there are one or more deferred tokens, load the current row into
** memory and scan it to determine the position list for each deferred
** token. Then, see if this row is really a match, considering deferred
** tokens and NEAR operators (neither of which were taken into account
** earlier, by fts3EvalNextRow()).
*/
if( pCsr->pDeferred ){
rc = fts3CursorSeek(0, pCsr);
if( rc==SQLITE_OK ){
rc = sqlite3Fts3CacheDeferredDoclists(pCsr);
}
}
bMiss = (0==fts3EvalTestExpr(pCsr, pCsr->pExpr, &rc));
/* Free the position-lists accumulated for each deferred token above. */
sqlite3Fts3FreeDeferredDoclists(pCsr);
*pRc = rc;
}
return (rc==SQLITE_OK && bMiss);
}
/*
** Advance to the next document that matches the FTS expression in
** Fts3Cursor.pExpr.
*/
static int fts3EvalNext(Fts3Cursor *pCsr){
int rc = SQLITE_OK; /* Return Code */
Fts3Expr *pExpr = pCsr->pExpr;
assert( pCsr->isEof==0 );
if( pExpr==0 ){
pCsr->isEof = 1;
}else{
do {
if( pCsr->isRequireSeek==0 ){
sqlite3_reset(pCsr->pStmt);
}
assert( sqlite3_data_count(pCsr->pStmt)==0 );
fts3EvalNextRow(pCsr, pExpr, &rc);
pCsr->isEof = pExpr->bEof;
pCsr->isRequireSeek = 1;
pCsr->isMatchinfoNeeded = 1;
pCsr->iPrevId = pExpr->iDocid;
}while( pCsr->isEof==0 && sqlite3Fts3EvalTestDeferred(pCsr, &rc) );
}
/* Check if the cursor is past the end of the docid range specified
** by Fts3Cursor.iMinDocid/iMaxDocid. If so, set the EOF flag. */
if( rc==SQLITE_OK && (
(pCsr->bDesc==0 && pCsr->iPrevId>pCsr->iMaxDocid)
|| (pCsr->bDesc!=0 && pCsr->iPrevId<pCsr->iMinDocid)
)){
pCsr->isEof = 1;
}
return rc;
}
/*
** Restart interation for expression pExpr so that the next call to
** fts3EvalNext() visits the first row. Do not allow incremental
** loading or merging of phrase doclists for this iteration.
**
** If *pRc is other than SQLITE_OK when this function is called, it is
** a no-op. If an error occurs within this function, *pRc is set to an
** SQLite error code before returning.
*/
static void fts3EvalRestart(
Fts3Cursor *pCsr,
Fts3Expr *pExpr,
int *pRc
){
if( pExpr && *pRc==SQLITE_OK ){
Fts3Phrase *pPhrase = pExpr->pPhrase;
if( pPhrase ){
fts3EvalInvalidatePoslist(pPhrase);
if( pPhrase->bIncr ){
int i;
for(i=0; i<pPhrase->nToken; i++){
Fts3PhraseToken *pToken = &pPhrase->aToken[i];
assert( pToken->pDeferred==0 );
if( pToken->pSegcsr ){
sqlite3Fts3MsrIncrRestart(pToken->pSegcsr);
}
}
*pRc = fts3EvalPhraseStart(pCsr, 0, pPhrase);
}
pPhrase->doclist.pNextDocid = 0;
pPhrase->doclist.iDocid = 0;
pPhrase->pOrPoslist = 0;
}
pExpr->iDocid = 0;
pExpr->bEof = 0;
pExpr->bStart = 0;
fts3EvalRestart(pCsr, pExpr->pLeft, pRc);
fts3EvalRestart(pCsr, pExpr->pRight, pRc);
}
}
/*
** After allocating the Fts3Expr.aMI[] array for each phrase in the
** expression rooted at pExpr, the cursor iterates through all rows matched
** by pExpr, calling this function for each row. This function increments
** the values in Fts3Expr.aMI[] according to the position-list currently
** found in Fts3Expr.pPhrase->doclist.pList for each of the phrase
** expression nodes.
*/
static void fts3EvalUpdateCounts(Fts3Expr *pExpr, int nCol){
if( pExpr ){
Fts3Phrase *pPhrase = pExpr->pPhrase;
if( pPhrase && pPhrase->doclist.pList ){
int iCol = 0;
char *p = pPhrase->doclist.pList;
do{
u8 c = 0;
int iCnt = 0;
while( 0xFE & (*p | c) ){
if( (c&0x80)==0 ) iCnt++;
c = *p++ & 0x80;
}
/* aMI[iCol*3 + 1] = Number of occurrences
** aMI[iCol*3 + 2] = Number of rows containing at least one instance
*/
pExpr->aMI[iCol*3 + 1] += iCnt;
pExpr->aMI[iCol*3 + 2] += (iCnt>0);
if( *p==0x00 ) break;
p++;
p += fts3GetVarint32(p, &iCol);
}while( iCol<nCol );
}
fts3EvalUpdateCounts(pExpr->pLeft, nCol);
fts3EvalUpdateCounts(pExpr->pRight, nCol);
}
}
/*
** This is an sqlite3Fts3ExprIterate() callback. If the Fts3Expr.aMI[] array
** has not yet been allocated, allocate and zero it. Otherwise, just zero
** it.
*/
static int fts3AllocateMSI(Fts3Expr *pExpr, int iPhrase, void *pCtx){
Fts3Table *pTab = (Fts3Table*)pCtx;
UNUSED_PARAMETER(iPhrase);
if( pExpr->aMI==0 ){
pExpr->aMI = (u32 *)sqlite3_malloc64(pTab->nColumn * 3 * sizeof(u32));
if( pExpr->aMI==0 ) return SQLITE_NOMEM;
}
memset(pExpr->aMI, 0, pTab->nColumn * 3 * sizeof(u32));
return SQLITE_OK;
}
/*
** Expression pExpr must be of type FTSQUERY_PHRASE.
**
** If it is not already allocated and populated, this function allocates and
** populates the Fts3Expr.aMI[] array for expression pExpr. If pExpr is part
** of a NEAR expression, then it also allocates and populates the same array
** for all other phrases that are part of the NEAR expression.
**
** SQLITE_OK is returned if the aMI[] array is successfully allocated and
** populated. Otherwise, if an error occurs, an SQLite error code is returned.
*/
static int fts3EvalGatherStats(
Fts3Cursor *pCsr, /* Cursor object */
Fts3Expr *pExpr /* FTSQUERY_PHRASE expression */
){
int rc = SQLITE_OK; /* Return code */
assert( pExpr->eType==FTSQUERY_PHRASE );
if( pExpr->aMI==0 ){
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
Fts3Expr *pRoot; /* Root of NEAR expression */
sqlite3_int64 iPrevId = pCsr->iPrevId;
sqlite3_int64 iDocid;
u8 bEof;
/* Find the root of the NEAR expression */
pRoot = pExpr;
while( pRoot->pParent
&& (pRoot->pParent->eType==FTSQUERY_NEAR || pRoot->bDeferred)
){
pRoot = pRoot->pParent;
}
iDocid = pRoot->iDocid;
bEof = pRoot->bEof;
assert( pRoot->bStart );
/* Allocate space for the aMSI[] array of each FTSQUERY_PHRASE node */
rc = sqlite3Fts3ExprIterate(pRoot, fts3AllocateMSI, (void*)pTab);
if( rc!=SQLITE_OK ) return rc;
fts3EvalRestart(pCsr, pRoot, &rc);
while( pCsr->isEof==0 && rc==SQLITE_OK ){
do {
/* Ensure the %_content statement is reset. */
if( pCsr->isRequireSeek==0 ) sqlite3_reset(pCsr->pStmt);
assert( sqlite3_data_count(pCsr->pStmt)==0 );
/* Advance to the next document */
fts3EvalNextRow(pCsr, pRoot, &rc);
pCsr->isEof = pRoot->bEof;
pCsr->isRequireSeek = 1;
pCsr->isMatchinfoNeeded = 1;
pCsr->iPrevId = pRoot->iDocid;
}while( pCsr->isEof==0
&& pRoot->eType==FTSQUERY_NEAR
&& sqlite3Fts3EvalTestDeferred(pCsr, &rc)
);
if( rc==SQLITE_OK && pCsr->isEof==0 ){
fts3EvalUpdateCounts(pRoot, pTab->nColumn);
}
}
pCsr->isEof = 0;
pCsr->iPrevId = iPrevId;
if( bEof ){
pRoot->bEof = bEof;
}else{
/* Caution: pRoot may iterate through docids in ascending or descending
** order. For this reason, even though it seems more defensive, the
** do loop can not be written:
**
** do {...} while( pRoot->iDocid<iDocid && rc==SQLITE_OK );
*/
fts3EvalRestart(pCsr, pRoot, &rc);
do {
fts3EvalNextRow(pCsr, pRoot, &rc);
assert_fts3_nc( pRoot->bEof==0 );
if( pRoot->bEof ) rc = FTS_CORRUPT_VTAB;
}while( pRoot->iDocid!=iDocid && rc==SQLITE_OK );
}
}
return rc;
}
/*
** This function is used by the matchinfo() module to query a phrase
** expression node for the following information:
**
** 1. The total number of occurrences of the phrase in each column of
** the FTS table (considering all rows), and
**
** 2. For each column, the number of rows in the table for which the
** column contains at least one instance of the phrase.
**
** If no error occurs, SQLITE_OK is returned and the values for each column
** written into the array aiOut as follows:
**
** aiOut[iCol*3 + 1] = Number of occurrences
** aiOut[iCol*3 + 2] = Number of rows containing at least one instance
**
** Caveats:
**
** * If a phrase consists entirely of deferred tokens, then all output
** values are set to the number of documents in the table. In other
** words we assume that very common tokens occur exactly once in each
** column of each row of the table.
**
** * If a phrase contains some deferred tokens (and some non-deferred
** tokens), count the potential occurrence identified by considering
** the non-deferred tokens instead of actual phrase occurrences.
**
** * If the phrase is part of a NEAR expression, then only phrase instances
** that meet the NEAR constraint are included in the counts.
*/
int sqlite3Fts3EvalPhraseStats(
Fts3Cursor *pCsr, /* FTS cursor handle */
Fts3Expr *pExpr, /* Phrase expression */
u32 *aiOut /* Array to write results into (see above) */
){
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
int rc = SQLITE_OK;
int iCol;
if( pExpr->bDeferred && pExpr->pParent->eType!=FTSQUERY_NEAR ){
assert( pCsr->nDoc>0 );
for(iCol=0; iCol<pTab->nColumn; iCol++){
aiOut[iCol*3 + 1] = (u32)pCsr->nDoc;
aiOut[iCol*3 + 2] = (u32)pCsr->nDoc;
}
}else{
rc = fts3EvalGatherStats(pCsr, pExpr);
if( rc==SQLITE_OK ){
assert( pExpr->aMI );
for(iCol=0; iCol<pTab->nColumn; iCol++){
aiOut[iCol*3 + 1] = pExpr->aMI[iCol*3 + 1];
aiOut[iCol*3 + 2] = pExpr->aMI[iCol*3 + 2];
}
}
}
return rc;
}
/*
** The expression pExpr passed as the second argument to this function
** must be of type FTSQUERY_PHRASE.
**
** The returned value is either NULL or a pointer to a buffer containing
** a position-list indicating the occurrences of the phrase in column iCol
** of the current row.
**
** More specifically, the returned buffer contains 1 varint for each
** occurrence of the phrase in the column, stored using the normal (delta+2)
** compression and is terminated by either an 0x01 or 0x00 byte. For example,
** if the requested column contains "a b X c d X X" and the position-list
** for 'X' is requested, the buffer returned may contain:
**
** 0x04 0x05 0x03 0x01 or 0x04 0x05 0x03 0x00
**
** This function works regardless of whether or not the phrase is deferred,
** incremental, or neither.
*/
int sqlite3Fts3EvalPhrasePoslist(
Fts3Cursor *pCsr, /* FTS3 cursor object */
Fts3Expr *pExpr, /* Phrase to return doclist for */
int iCol, /* Column to return position list for */
char **ppOut /* OUT: Pointer to position list */
){
Fts3Phrase *pPhrase = pExpr->pPhrase;
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
char *pIter;
int iThis;
sqlite3_int64 iDocid;
/* If this phrase is applies specifically to some column other than
** column iCol, return a NULL pointer. */
*ppOut = 0;
assert( iCol>=0 && iCol<pTab->nColumn );
if( (pPhrase->iColumn<pTab->nColumn && pPhrase->iColumn!=iCol) ){
return SQLITE_OK;
}
iDocid = pExpr->iDocid;
pIter = pPhrase->doclist.pList;
if( iDocid!=pCsr->iPrevId || pExpr->bEof ){
int rc = SQLITE_OK;
int bDescDoclist = pTab->bDescIdx; /* For DOCID_CMP macro */
int bOr = 0;
u8 bTreeEof = 0;
Fts3Expr *p; /* Used to iterate from pExpr to root */
Fts3Expr *pNear; /* Most senior NEAR ancestor (or pExpr) */
Fts3Expr *pRun; /* Closest non-deferred ancestor of pNear */
int bMatch;
/* Check if this phrase descends from an OR expression node. If not,
** return NULL. Otherwise, the entry that corresponds to docid
** pCsr->iPrevId may lie earlier in the doclist buffer. Or, if the
** tree that the node is part of has been marked as EOF, but the node
** itself is not EOF, then it may point to an earlier entry. */
pNear = pExpr;
for(p=pExpr->pParent; p; p=p->pParent){
if( p->eType==FTSQUERY_OR ) bOr = 1;
if( p->eType==FTSQUERY_NEAR ) pNear = p;
if( p->bEof ) bTreeEof = 1;
}
if( bOr==0 ) return SQLITE_OK;
pRun = pNear;
while( pRun->bDeferred ){
assert( pRun->pParent );
pRun = pRun->pParent;
}
/* This is the descendent of an OR node. In this case we cannot use
** an incremental phrase. Load the entire doclist for the phrase
** into memory in this case. */
if( pPhrase->bIncr ){
int bEofSave = pRun->bEof;
fts3EvalRestart(pCsr, pRun, &rc);
while( rc==SQLITE_OK && !pRun->bEof ){
fts3EvalNextRow(pCsr, pRun, &rc);
if( bEofSave==0 && pRun->iDocid==iDocid ) break;
}
assert( rc!=SQLITE_OK || pPhrase->bIncr==0 );
if( rc==SQLITE_OK && pRun->bEof!=bEofSave ){
rc = FTS_CORRUPT_VTAB;
}
}
if( bTreeEof ){
while( rc==SQLITE_OK && !pRun->bEof ){
fts3EvalNextRow(pCsr, pRun, &rc);
}
}
if( rc!=SQLITE_OK ) return rc;
bMatch = 1;
for(p=pNear; p; p=p->pLeft){
u8 bEof = 0;
Fts3Expr *pTest = p;
Fts3Phrase *pPh;
assert( pTest->eType==FTSQUERY_NEAR || pTest->eType==FTSQUERY_PHRASE );
if( pTest->eType==FTSQUERY_NEAR ) pTest = pTest->pRight;
assert( pTest->eType==FTSQUERY_PHRASE );
pPh = pTest->pPhrase;
pIter = pPh->pOrPoslist;
iDocid = pPh->iOrDocid;
if( pCsr->bDesc==bDescDoclist ){
bEof = !pPh->doclist.nAll ||
(pIter >= (pPh->doclist.aAll + pPh->doclist.nAll));
while( (pIter==0 || DOCID_CMP(iDocid, pCsr->iPrevId)<0 ) && bEof==0 ){
sqlite3Fts3DoclistNext(
bDescDoclist, pPh->doclist.aAll, pPh->doclist.nAll,
&pIter, &iDocid, &bEof
);
}
}else{
bEof = !pPh->doclist.nAll || (pIter && pIter<=pPh->doclist.aAll);
while( (pIter==0 || DOCID_CMP(iDocid, pCsr->iPrevId)>0 ) && bEof==0 ){
int dummy;
sqlite3Fts3DoclistPrev(
bDescDoclist, pPh->doclist.aAll, pPh->doclist.nAll,
&pIter, &iDocid, &dummy, &bEof
);
}
}
pPh->pOrPoslist = pIter;
pPh->iOrDocid = iDocid;
if( bEof || iDocid!=pCsr->iPrevId ) bMatch = 0;
}
if( bMatch ){
pIter = pPhrase->pOrPoslist;
}else{
pIter = 0;
}
}
if( pIter==0 ) return SQLITE_OK;
if( *pIter==0x01 ){
pIter++;
pIter += fts3GetVarint32(pIter, &iThis);
}else{
iThis = 0;
}
while( iThis<iCol ){
fts3ColumnlistCopy(0, &pIter);
if( *pIter==0x00 ) return SQLITE_OK;
pIter++;
pIter += fts3GetVarint32(pIter, &iThis);
}
if( *pIter==0x00 ){
pIter = 0;
}
*ppOut = ((iCol==iThis)?pIter:0);
return SQLITE_OK;
}
/*
** Free all components of the Fts3Phrase structure that were allocated by
** the eval module. Specifically, this means to free:
**
** * the contents of pPhrase->doclist, and
** * any Fts3MultiSegReader objects held by phrase tokens.
*/
void sqlite3Fts3EvalPhraseCleanup(Fts3Phrase *pPhrase){
if( pPhrase ){
int i;
sqlite3_free(pPhrase->doclist.aAll);
fts3EvalInvalidatePoslist(pPhrase);
memset(&pPhrase->doclist, 0, sizeof(Fts3Doclist));
for(i=0; i<pPhrase->nToken; i++){
fts3SegReaderCursorFree(pPhrase->aToken[i].pSegcsr);
pPhrase->aToken[i].pSegcsr = 0;
}
}
}
/*
** Return SQLITE_CORRUPT_VTAB.
*/
#ifdef SQLITE_DEBUG
int sqlite3Fts3Corrupt(){
return SQLITE_CORRUPT_VTAB;
}
#endif
#if !SQLITE_CORE
/*
** Initialize API pointer table, if required.
*/
#ifdef _WIN32
__declspec(dllexport)
#endif
int sqlite3_fts3_init(
sqlite3 *db,
char **pzErrMsg,
const sqlite3_api_routines *pApi
){
SQLITE_EXTENSION_INIT2(pApi)
return sqlite3Fts3Init(db);
}
#endif
#endif
|