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
|
/* Copyright (c) 2000, 2014, Oracle and/or its affiliates.
Copyright (c) 2009, 2014, Monty Program Ab.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA */
/**
@file
@brief
Sorts a database
*/
#include <my_global.h>
#include "sql_priv.h"
#include "filesort.h"
#include "unireg.h" // REQUIRED by other includes
#ifdef HAVE_STDDEF_H
#include <stddef.h> /* for macro offsetof */
#endif
#include <m_ctype.h>
#include "sql_sort.h"
#include "probes_mysql.h"
#include "sql_base.h" // update_virtual_fields
#include "sql_test.h" // TEST_filesort
#include "opt_range.h" // SQL_SELECT
#include "bounded_queue.h"
#include "filesort_utils.h"
#include "sql_select.h"
#include "log_slow.h"
#include "debug_sync.h"
/// How to write record_ref.
#define WRITE_REF(file,from) \
if (my_b_write((file),(uchar*) (from),param->ref_length)) \
DBUG_RETURN(1);
/* functions defined in this file */
static uchar *read_buffpek_from_file(IO_CACHE *buffer_file, uint count,
uchar *buf);
static ha_rows find_all_keys(Sort_param *param,SQL_SELECT *select,
Filesort_info *fs_info,
IO_CACHE *buffer_file,
IO_CACHE *tempfile,
Bounded_queue<uchar, uchar> *pq,
ha_rows *found_rows);
static bool write_keys(Sort_param *param, Filesort_info *fs_info,
uint count, IO_CACHE *buffer_file, IO_CACHE *tempfile);
static void make_sortkey(Sort_param *param, uchar *to, uchar *ref_pos);
static void register_used_fields(Sort_param *param);
static bool save_index(Sort_param *param, uint count,
Filesort_info *table_sort);
static uint suffix_length(ulong string_length);
static uint sortlength(THD *thd, SORT_FIELD *sortorder, uint s_length,
bool *multi_byte_charset);
static SORT_ADDON_FIELD *get_addon_fields(ulong max_length_for_sort_data,
Field **ptabfield,
uint sortlength, uint *plength);
static void unpack_addon_fields(struct st_sort_addon_field *addon_field,
uchar *buff, uchar *buff_end);
static bool check_if_pq_applicable(Sort_param *param, Filesort_info *info,
TABLE *table,
ha_rows records, ulong memory_available);
void Sort_param::init_for_filesort(uint sortlen, TABLE *table,
ulong max_length_for_sort_data,
ha_rows maxrows, bool sort_positions)
{
sort_length= sortlen;
ref_length= table->file->ref_length;
if (!(table->file->ha_table_flags() & HA_FAST_KEY_READ) &&
!table->fulltext_searched && !sort_positions)
{
/*
Get the descriptors of all fields whose values are appended
to sorted fields and get its total length in addon_length.
*/
addon_field= get_addon_fields(max_length_for_sort_data,
table->field, sort_length, &addon_length);
}
if (addon_field)
res_length= addon_length;
else
{
res_length= ref_length;
/*
The reference to the record is considered
as an additional sorted field
*/
sort_length+= ref_length;
}
rec_length= sort_length + addon_length;
max_rows= maxrows;
}
/**
Sort a table.
Creates a set of pointers that can be used to read the rows
in sorted order. This should be done with the functions
in records.cc.
Before calling filesort, one must have done
table->file->info(HA_STATUS_VARIABLE)
The result set is stored in table->io_cache or
table->record_pointers.
@param thd Current thread
@param table Table to sort
@param sortorder How to sort the table
@param s_length Number of elements in sortorder
@param select Condition to apply to the rows
@param max_rows Return only this many rows
@param sort_positions Set to TRUE if we want to force sorting by position
(Needed by UPDATE/INSERT or ALTER TABLE or
when rowids are required by executor)
@param[out] examined_rows Store number of examined rows here
@param[out] found_rows Store the number of found rows here
@note
If we sort by position (like if sort_positions is 1) filesort() will
call table->prepare_for_position().
@retval
HA_POS_ERROR Error
@retval
\# Number of rows
*/
ha_rows filesort(THD *thd, TABLE *table, SORT_FIELD *sortorder, uint s_length,
SQL_SELECT *select, ha_rows max_rows,
bool sort_positions,
ha_rows *examined_rows,
ha_rows *found_rows)
{
int error;
size_t memory_available= thd->variables.sortbuff_size;
uint maxbuffer;
BUFFPEK *buffpek;
ha_rows num_rows= HA_POS_ERROR;
IO_CACHE tempfile, buffpek_pointers, *outfile;
Sort_param param;
bool multi_byte_charset;
Bounded_queue<uchar, uchar> pq;
DBUG_ENTER("filesort");
DBUG_EXECUTE("info",TEST_filesort(sortorder,s_length););
#ifdef SKIP_DBUG_IN_FILESORT
DBUG_PUSH(""); /* No DBUG here */
#endif
Filesort_info table_sort= table->sort;
TABLE_LIST *tab= table->pos_in_table_list;
Item_subselect *subselect= tab ? tab->containing_subselect() : 0;
MYSQL_FILESORT_START(table->s->db.str, table->s->table_name.str);
DEBUG_SYNC(thd, "filesort_start");
/*
Release InnoDB's adaptive hash index latch (if holding) before
running a sort.
*/
ha_release_temporary_latches(thd);
/*
Don't use table->sort in filesort as it is also used by
QUICK_INDEX_MERGE_SELECT. Work with a copy and put it back at the end
when index_merge select has finished with it.
*/
table->sort.io_cache= NULL;
DBUG_ASSERT(table_sort.record_pointers == NULL);
outfile= table_sort.io_cache;
my_b_clear(&tempfile);
my_b_clear(&buffpek_pointers);
buffpek=0;
error= 1;
*found_rows= HA_POS_ERROR;
param.init_for_filesort(sortlength(thd, sortorder, s_length,
&multi_byte_charset),
table,
thd->variables.max_length_for_sort_data,
max_rows, sort_positions);
table_sort.addon_buf= 0;
table_sort.addon_length= param.addon_length;
table_sort.addon_field= param.addon_field;
table_sort.unpack= unpack_addon_fields;
if (param.addon_field &&
!(table_sort.addon_buf=
(uchar *) my_malloc(param.addon_length, MYF(MY_WME |
MY_THREAD_SPECIFIC))))
goto err;
if (select && select->quick)
thd->inc_status_sort_range();
else
thd->inc_status_sort_scan();
thd->query_plan_flags|= QPLAN_FILESORT;
// If number of rows is not known, use as much of sort buffer as possible.
num_rows= table->file->estimate_rows_upper_bound();
if (multi_byte_charset &&
!(param.tmp_buffer= (char*) my_malloc(param.sort_length,
MYF(MY_WME | MY_THREAD_SPECIFIC))))
goto err;
if (check_if_pq_applicable(¶m, &table_sort,
table, num_rows, memory_available))
{
DBUG_PRINT("info", ("filesort PQ is applicable"));
thd->query_plan_flags|= QPLAN_FILESORT_PRIORITY_QUEUE;
status_var_increment(thd->status_var.filesort_pq_sorts_);
const size_t compare_length= param.sort_length;
if (pq.init(param.max_rows,
true, // max_at_top
NULL, // compare_function
compare_length,
&make_sortkey, ¶m, table_sort.get_sort_keys()))
{
/*
If we fail to init pq, we have to give up:
out of memory means my_malloc() will call my_error().
*/
DBUG_PRINT("info", ("failed to allocate PQ"));
table_sort.free_sort_buffer();
DBUG_ASSERT(thd->is_error());
goto err;
}
// For PQ queries (with limit) we initialize all pointers.
table_sort.init_record_pointers();
}
else
{
DBUG_PRINT("info", ("filesort PQ is not applicable"));
size_t min_sort_memory= MY_MAX(MIN_SORT_MEMORY, param.sort_length*MERGEBUFF2);
set_if_bigger(min_sort_memory, sizeof(BUFFPEK*)*MERGEBUFF2);
while (memory_available >= min_sort_memory)
{
ulonglong keys= memory_available / (param.rec_length + sizeof(char*));
param.max_keys_per_buffer= (uint) MY_MIN(num_rows, keys);
if (table_sort.get_sort_keys())
{
// If we have already allocated a buffer, it better have same size!
if (!table_sort.check_sort_buffer_properties(param.max_keys_per_buffer,
param.rec_length))
{
/*
table->sort will still have a pointer to the same buffer,
but that will be overwritten by the assignment below.
*/
table_sort.free_sort_buffer();
}
}
table_sort.alloc_sort_buffer(param.max_keys_per_buffer, param.rec_length);
if (table_sort.get_sort_keys())
break;
size_t old_memory_available= memory_available;
memory_available= memory_available/4*3;
if (memory_available < min_sort_memory &&
old_memory_available > min_sort_memory)
memory_available= min_sort_memory;
}
if (memory_available < min_sort_memory)
{
my_error(ER_OUT_OF_SORTMEMORY,MYF(ME_ERROR + ME_FATALERROR));
goto err;
}
}
if (open_cached_file(&buffpek_pointers,mysql_tmpdir,TEMP_PREFIX,
DISK_BUFFER_SIZE, MYF(MY_WME)))
goto err;
param.sort_form= table;
param.end=(param.local_sortorder=sortorder)+s_length;
num_rows= find_all_keys(¶m, select,
&table_sort,
&buffpek_pointers,
&tempfile,
pq.is_initialized() ? &pq : NULL,
found_rows);
if (num_rows == HA_POS_ERROR)
goto err;
maxbuffer= (uint) (my_b_tell(&buffpek_pointers)/sizeof(*buffpek));
if (maxbuffer == 0) // The whole set is in memory
{
if (save_index(¶m, (uint) num_rows, &table_sort))
goto err;
}
else
{
/* filesort cannot handle zero-length records during merge. */
DBUG_ASSERT(param.sort_length != 0);
if (table_sort.buffpek && table_sort.buffpek_len < maxbuffer)
{
my_free(table_sort.buffpek);
table_sort.buffpek= 0;
}
if (!(table_sort.buffpek=
(uchar *) read_buffpek_from_file(&buffpek_pointers, maxbuffer,
table_sort.buffpek)))
goto err;
buffpek= (BUFFPEK *) table_sort.buffpek;
table_sort.buffpek_len= maxbuffer;
close_cached_file(&buffpek_pointers);
/* Open cached file if it isn't open */
if (! my_b_inited(outfile) &&
open_cached_file(outfile,mysql_tmpdir,TEMP_PREFIX,READ_RECORD_BUFFER,
MYF(MY_WME)))
goto err;
if (reinit_io_cache(outfile,WRITE_CACHE,0L,0,0))
goto err;
/*
Use also the space previously used by string pointers in sort_buffer
for temporary key storage.
*/
param.max_keys_per_buffer=((param.max_keys_per_buffer *
(param.rec_length + sizeof(char*))) /
param.rec_length - 1);
maxbuffer--; // Offset from 0
if (merge_many_buff(¶m,
(uchar*) table_sort.get_sort_keys(),
buffpek,&maxbuffer,
&tempfile))
goto err;
if (flush_io_cache(&tempfile) ||
reinit_io_cache(&tempfile,READ_CACHE,0L,0,0))
goto err;
if (merge_index(¶m,
(uchar*) table_sort.get_sort_keys(),
buffpek,
maxbuffer,
&tempfile,
outfile))
goto err;
}
if (num_rows > param.max_rows)
{
// If find_all_keys() produced more results than the query LIMIT.
num_rows= param.max_rows;
}
error= 0;
err:
my_free(param.tmp_buffer);
if (!subselect || !subselect->is_uncacheable())
{
table_sort.free_sort_buffer();
my_free(buffpek);
table_sort.buffpek= 0;
table_sort.buffpek_len= 0;
}
close_cached_file(&tempfile);
close_cached_file(&buffpek_pointers);
if (my_b_inited(outfile))
{
if (flush_io_cache(outfile))
error=1;
{
my_off_t save_pos=outfile->pos_in_file;
/* For following reads */
if (reinit_io_cache(outfile,READ_CACHE,0L,0,0))
error=1;
outfile->end_of_file=save_pos;
}
}
if (error)
{
int kill_errno= thd->killed_errno();
DBUG_ASSERT(thd->is_error() || kill_errno || thd->killed == ABORT_QUERY);
/*
We replace the table->sort at the end.
Hence calling free_io_cache to make sure table->sort.io_cache
used for QUICK_INDEX_MERGE_SELECT is free.
*/
free_io_cache(table);
my_printf_error(ER_FILSORT_ABORT,
"%s: %s",
MYF(0),
ER_THD(thd, ER_FILSORT_ABORT),
kill_errno ? ER(kill_errno) :
thd->killed == ABORT_QUERY ? "" :
thd->get_stmt_da()->message());
if (global_system_variables.log_warnings > 1)
{
sql_print_warning("%s, host: %s, user: %s, thread: %lu, query: %-.4096s",
ER_THD(thd, ER_FILSORT_ABORT),
thd->security_ctx->host_or_ip,
&thd->security_ctx->priv_user[0],
(ulong) thd->thread_id,
thd->query());
}
}
else
thd->inc_status_sort_rows(num_rows);
*examined_rows= param.examined_rows;
#ifdef SKIP_DBUG_IN_FILESORT
DBUG_POP(); /* Ok to DBUG */
#endif
/* table->sort.io_cache should be free by this time */
DBUG_ASSERT(NULL == table->sort.io_cache);
// Assign the copy back!
table->sort= table_sort;
DBUG_PRINT("exit",
("num_rows: %ld examined_rows: %ld found_rows: %ld",
(long) num_rows, (long) *examined_rows, (long) *found_rows));
MYSQL_FILESORT_DONE(error, num_rows);
DBUG_RETURN(error ? HA_POS_ERROR : num_rows);
} /* filesort */
void filesort_free_buffers(TABLE *table, bool full)
{
DBUG_ENTER("filesort_free_buffers");
my_free(table->sort.record_pointers);
table->sort.record_pointers= NULL;
if (full)
{
table->sort.free_sort_buffer();
my_free(table->sort.buffpek);
table->sort.buffpek= NULL;
table->sort.buffpek_len= 0;
}
my_free(table->sort.addon_buf);
my_free(table->sort.addon_field);
table->sort.addon_buf= NULL;
table->sort.addon_field= NULL;
DBUG_VOID_RETURN;
}
/** Read 'count' number of buffer pointers into memory. */
static uchar *read_buffpek_from_file(IO_CACHE *buffpek_pointers, uint count,
uchar *buf)
{
size_t length= sizeof(BUFFPEK)*count;
uchar *tmp= buf;
DBUG_ENTER("read_buffpek_from_file");
if (count > UINT_MAX/sizeof(BUFFPEK))
return 0; /* sizeof(BUFFPEK)*count will overflow */
if (!tmp)
tmp= (uchar *)my_malloc(length, MYF(MY_WME | MY_THREAD_SPECIFIC));
if (tmp)
{
if (reinit_io_cache(buffpek_pointers,READ_CACHE,0L,0,0) ||
my_b_read(buffpek_pointers, (uchar*) tmp, length))
{
my_free(tmp);
tmp=0;
}
}
DBUG_RETURN(tmp);
}
#ifndef DBUG_OFF
/* Buffer where record is returned */
char dbug_print_row_buff[512];
/* Temporary buffer for printing a column */
char dbug_print_row_buff_tmp[512];
/*
Print table's current row into a buffer and return a pointer to it.
This is intended to be used from gdb:
(gdb) p dbug_print_table_row(table)
$33 = "SUBQUERY2_t1(col_int_key,col_varchar_nokey)=(7,c)"
(gdb)
Only columns in table->read_set are printed
*/
const char* dbug_print_table_row(TABLE *table)
{
Field **pfield;
String tmp(dbug_print_row_buff_tmp,
sizeof(dbug_print_row_buff_tmp),&my_charset_bin);
String output(dbug_print_row_buff, sizeof(dbug_print_row_buff),
&my_charset_bin);
output.length(0);
output.append(table->alias);
output.append("(");
bool first= true;
for (pfield= table->field; *pfield ; pfield++)
{
if (table->read_set && !bitmap_is_set(table->read_set, (*pfield)->field_index))
continue;
if (first)
first= false;
else
output.append(",");
output.append((*pfield)->field_name? (*pfield)->field_name: "NULL");
}
output.append(")=(");
first= true;
for (pfield= table->field; *pfield ; pfield++)
{
Field *field= *pfield;
if (table->read_set && !bitmap_is_set(table->read_set, (*pfield)->field_index))
continue;
if (first)
first= false;
else
output.append(",");
if (field->is_null())
output.append("NULL");
else
{
if (field->type() == MYSQL_TYPE_BIT)
(void) field->val_int_as_str(&tmp, 1);
else
field->val_str(&tmp);
output.append(tmp.ptr(), tmp.length());
}
}
output.append(")");
return output.c_ptr_safe();
}
/*
Print a text, SQL-like record representation into dbug trace.
Note: this function is a work in progress: at the moment
- column read bitmap is ignored (can print garbage for unused columns)
- there is no quoting
*/
static void dbug_print_record(TABLE *table, bool print_rowid)
{
char buff[1024];
Field **pfield;
String tmp(buff,sizeof(buff),&my_charset_bin);
DBUG_LOCK_FILE;
fprintf(DBUG_FILE, "record (");
for (pfield= table->field; *pfield ; pfield++)
fprintf(DBUG_FILE, "%s%s", (*pfield)->field_name, (pfield[1])? ", ":"");
fprintf(DBUG_FILE, ") = ");
fprintf(DBUG_FILE, "(");
for (pfield= table->field; *pfield ; pfield++)
{
Field *field= *pfield;
if (field->is_null())
fwrite("NULL", sizeof(char), 4, DBUG_FILE);
if (field->type() == MYSQL_TYPE_BIT)
(void) field->val_int_as_str(&tmp, 1);
else
field->val_str(&tmp);
fwrite(tmp.ptr(),sizeof(char),tmp.length(),DBUG_FILE);
if (pfield[1])
fwrite(", ", sizeof(char), 2, DBUG_FILE);
}
fprintf(DBUG_FILE, ")");
if (print_rowid)
{
fprintf(DBUG_FILE, " rowid ");
for (uint i=0; i < table->file->ref_length; i++)
{
fprintf(DBUG_FILE, "%x", (uchar)table->file->ref[i]);
}
}
fprintf(DBUG_FILE, "\n");
DBUG_UNLOCK_FILE;
}
#endif
/**
Search after sort_keys, and write them into tempfile
(if we run out of space in the sort_keys buffer).
All produced sequences are guaranteed to be non-empty.
@param param Sorting parameter
@param select Use this to get source data
@param sort_keys Array of pointers to sort key + addon buffers.
@param buffpek_pointers File to write BUFFPEKs describing sorted segments
in tempfile.
@param tempfile File to write sorted sequences of sortkeys to.
@param pq If !NULL, use it for keeping top N elements
@param [out] found_rows The number of FOUND_ROWS().
For a query with LIMIT, this value will typically
be larger than the function return value.
@note
Basic idea:
@verbatim
while (get_next_sortkey())
{
if (using priority queue)
push sort key into queue
else
{
if (no free space in sort_keys buffers)
{
sort sort_keys buffer;
dump sorted sequence to 'tempfile';
dump BUFFPEK describing sequence location into 'buffpek_pointers';
}
put sort key into 'sort_keys';
}
}
if (sort_keys has some elements && dumped at least once)
sort-dump-dump as above;
else
don't sort, leave sort_keys array to be sorted by caller.
@endverbatim
@retval
Number of records written on success.
@retval
HA_POS_ERROR on error.
*/
static ha_rows find_all_keys(Sort_param *param, SQL_SELECT *select,
Filesort_info *fs_info,
IO_CACHE *buffpek_pointers,
IO_CACHE *tempfile,
Bounded_queue<uchar, uchar> *pq,
ha_rows *found_rows)
{
int error,flag,quick_select;
uint idx,indexpos,ref_length;
uchar *ref_pos,*next_pos,ref_buff[MAX_REFLENGTH];
my_off_t record;
TABLE *sort_form;
THD *thd= current_thd;
handler *file;
MY_BITMAP *save_read_set, *save_write_set, *save_vcol_set;
DBUG_ENTER("find_all_keys");
DBUG_PRINT("info",("using: %s",
(select ? select->quick ? "ranges" : "where":
"every row")));
idx=indexpos=0;
error=quick_select=0;
sort_form=param->sort_form;
file=sort_form->file;
ref_length=param->ref_length;
ref_pos= ref_buff;
quick_select=select && select->quick;
record=0;
*found_rows= 0;
flag= ((file->ha_table_flags() & HA_REC_NOT_IN_SEQ) || quick_select);
if (flag)
ref_pos= &file->ref[0];
next_pos=ref_pos;
DBUG_EXECUTE_IF("show_explain_in_find_all_keys",
dbug_serve_apcs(thd, 1);
);
if (!quick_select)
{
next_pos=(uchar*) 0; /* Find records in sequence */
DBUG_EXECUTE_IF("bug14365043_1",
DBUG_SET("+d,ha_rnd_init_fail"););
if (file->ha_rnd_init_with_error(1))
DBUG_RETURN(HA_POS_ERROR);
file->extra_opt(HA_EXTRA_CACHE,
current_thd->variables.read_buff_size);
}
/* Remember original bitmaps */
save_read_set= sort_form->read_set;
save_write_set= sort_form->write_set;
save_vcol_set= sort_form->vcol_set;
/* Set up temporary column read map for columns used by sort */
bitmap_clear_all(&sort_form->tmp_set);
/* Temporary set for register_used_fields and register_field_in_read_map */
sort_form->read_set= &sort_form->tmp_set;
register_used_fields(param);
if (quick_select)
select->quick->add_used_key_part_to_set(sort_form->read_set);
Item *sort_cond= !select ?
0 : !select->pre_idx_push_select_cond ?
select->cond : select->pre_idx_push_select_cond;
if (sort_cond)
sort_cond->walk(&Item::register_field_in_read_map, 1, (uchar*) sort_form);
sort_form->column_bitmaps_set(&sort_form->tmp_set, &sort_form->tmp_set,
&sort_form->tmp_set);
if (quick_select)
{
if (select->quick->reset())
DBUG_RETURN(HA_POS_ERROR);
}
DEBUG_SYNC(thd, "after_index_merge_phase1");
for (;;)
{
if (quick_select)
{
if ((error= select->quick->get_next()))
break;
if (!error && sort_form->vfield)
update_virtual_fields(thd, sort_form);
file->position(sort_form->record[0]);
DBUG_EXECUTE_IF("debug_filesort", dbug_print_record(sort_form, TRUE););
}
else /* Not quick-select */
{
{
error= file->ha_rnd_next(sort_form->record[0]);
if (!error && sort_form->vfield)
update_virtual_fields(thd, sort_form);
if (!flag)
{
my_store_ptr(ref_pos,ref_length,record); // Position to row
record+= sort_form->s->db_record_offset;
}
else if (!error)
file->position(sort_form->record[0]);
}
if (error && error != HA_ERR_RECORD_DELETED)
break;
}
if (thd->check_killed())
{
DBUG_PRINT("info",("Sort killed by user"));
if (!quick_select)
{
(void) file->extra(HA_EXTRA_NO_CACHE);
file->ha_rnd_end();
}
DBUG_RETURN(HA_POS_ERROR); /* purecov: inspected */
}
bool write_record= false;
if (error == 0)
{
param->examined_rows++;
if (select && select->cond)
{
/*
If the condition 'select->cond' contains a subquery, restore the
original read/write sets of the table 'sort_form' because when
SQL_SELECT::skip_record evaluates this condition. it may include a
correlated subquery predicate, such that some field in the subquery
refers to 'sort_form'.
PSergey-todo: discuss the above with Timour.
*/
MY_BITMAP *tmp_read_set= sort_form->read_set;
MY_BITMAP *tmp_write_set= sort_form->write_set;
MY_BITMAP *tmp_vcol_set= sort_form->vcol_set;
if (select->cond->with_subselect)
sort_form->column_bitmaps_set(save_read_set, save_write_set,
save_vcol_set);
write_record= (select->skip_record(thd) > 0);
if (select->cond->with_subselect)
sort_form->column_bitmaps_set(tmp_read_set,
tmp_write_set,
tmp_vcol_set);
}
else
write_record= true;
}
if (write_record)
{
++(*found_rows);
if (pq)
{
pq->push(ref_pos);
idx= pq->num_elements();
}
else
{
if (idx == param->max_keys_per_buffer)
{
if (write_keys(param, fs_info, idx, buffpek_pointers, tempfile))
DBUG_RETURN(HA_POS_ERROR);
idx= 0;
indexpos++;
}
make_sortkey(param, fs_info->get_record_buffer(idx++), ref_pos);
}
}
/* It does not make sense to read more keys in case of a fatal error */
if (thd->is_error())
break;
/*
We need to this after checking the error as the transaction may have
rolled back in case of a deadlock
*/
if (!write_record)
file->unlock_row();
}
if (!quick_select)
{
(void) file->extra(HA_EXTRA_NO_CACHE); /* End cacheing of records */
if (!next_pos)
file->ha_rnd_end();
}
if (thd->is_error())
DBUG_RETURN(HA_POS_ERROR);
/* Signal we should use orignal column read and write maps */
sort_form->column_bitmaps_set(save_read_set, save_write_set, save_vcol_set);
DBUG_PRINT("test",("error: %d indexpos: %d",error,indexpos));
if (error != HA_ERR_END_OF_FILE)
{
file->print_error(error,MYF(ME_ERROR | ME_WAITTANG)); // purecov: inspected
DBUG_RETURN(HA_POS_ERROR); /* purecov: inspected */
}
if (indexpos && idx &&
write_keys(param, fs_info, idx, buffpek_pointers, tempfile))
DBUG_RETURN(HA_POS_ERROR); /* purecov: inspected */
const ha_rows retval=
my_b_inited(tempfile) ?
(ha_rows) (my_b_tell(tempfile)/param->rec_length) : idx;
DBUG_PRINT("info", ("find_all_keys return %u", (uint) retval));
DBUG_RETURN(retval);
} /* find_all_keys */
/**
@details
Sort the buffer and write:
-# the sorted sequence to tempfile
-# a BUFFPEK describing the sorted sequence position to buffpek_pointers
(was: Skriver en buffert med nycklar till filen)
@param param Sort parameters
@param sort_keys Array of pointers to keys to sort
@param count Number of elements in sort_keys array
@param buffpek_pointers One 'BUFFPEK' struct will be written into this file.
The BUFFPEK::{file_pos, count} will indicate where
the sorted data was stored.
@param tempfile The sorted sequence will be written into this file.
@retval
0 OK
@retval
1 Error
*/
static bool
write_keys(Sort_param *param, Filesort_info *fs_info, uint count,
IO_CACHE *buffpek_pointers, IO_CACHE *tempfile)
{
size_t rec_length;
uchar **end;
BUFFPEK buffpek;
DBUG_ENTER("write_keys");
rec_length= param->rec_length;
uchar **sort_keys= fs_info->get_sort_keys();
fs_info->sort_buffer(param, count);
if (!my_b_inited(tempfile) &&
open_cached_file(tempfile, mysql_tmpdir, TEMP_PREFIX, DISK_BUFFER_SIZE,
MYF(MY_WME)))
goto err; /* purecov: inspected */
/* check we won't have more buffpeks than we can possibly keep in memory */
if (my_b_tell(buffpek_pointers) + sizeof(BUFFPEK) > (ulonglong)UINT_MAX)
goto err;
buffpek.file_pos= my_b_tell(tempfile);
if ((ha_rows) count > param->max_rows)
count=(uint) param->max_rows; /* purecov: inspected */
buffpek.count=(ha_rows) count;
for (end=sort_keys+count ; sort_keys != end ; sort_keys++)
if (my_b_write(tempfile, (uchar*) *sort_keys, (uint) rec_length))
goto err;
if (my_b_write(buffpek_pointers, (uchar*) &buffpek, sizeof(buffpek)))
goto err;
DBUG_RETURN(0);
err:
DBUG_RETURN(1);
} /* write_keys */
/**
Store length as suffix in high-byte-first order.
*/
static inline void store_length(uchar *to, uint length, uint pack_length)
{
switch (pack_length) {
case 1:
*to= (uchar) length;
break;
case 2:
mi_int2store(to, length);
break;
case 3:
mi_int3store(to, length);
break;
default:
mi_int4store(to, length);
break;
}
}
/** Make a sort-key from record. */
static void make_sortkey(register Sort_param *param,
register uchar *to, uchar *ref_pos)
{
reg3 Field *field;
reg1 SORT_FIELD *sort_field;
reg5 uint length;
for (sort_field=param->local_sortorder ;
sort_field != param->end ;
sort_field++)
{
bool maybe_null=0;
if ((field=sort_field->field))
{ // Field
field->make_sort_key(to, sort_field->length);
if ((maybe_null = field->maybe_null()))
to++;
}
else
{ // Item
Item *item=sort_field->item;
maybe_null= item->maybe_null;
switch (sort_field->result_type) {
case STRING_RESULT:
{
const CHARSET_INFO *cs=item->collation.collation;
char fill_char= ((cs->state & MY_CS_BINSORT) ? (char) 0 : ' ');
if (maybe_null)
*to++=1;
char *tmp_buffer= param->tmp_buffer ? param->tmp_buffer : (char*)to;
String tmp(tmp_buffer, param->sort_length, cs);
String *res= item->str_result(&tmp);
if (!res)
{
if (maybe_null)
memset(to-1, 0, sort_field->length+1);
else
{
/* purecov: begin deadcode */
/*
This should only happen during extreme conditions if we run out
of memory or have an item marked not null when it can be null.
This code is here mainly to avoid a hard crash in this case.
*/
DBUG_ASSERT(0);
DBUG_PRINT("warning",
("Got null on something that shouldn't be null"));
memset(to, 0, sort_field->length); // Avoid crash
/* purecov: end */
}
break;
}
length= res->length();
if (sort_field->need_strxnfrm)
{
uint tmp_length __attribute__((unused));
tmp_length= cs->coll->strnxfrm(cs, to, sort_field->length,
item->max_char_length() *
cs->strxfrm_multiply,
(uchar*) res->ptr(), length,
MY_STRXFRM_PAD_WITH_SPACE |
MY_STRXFRM_PAD_TO_MAXLEN);
DBUG_ASSERT(tmp_length == sort_field->length);
}
else
{
uint diff;
uint sort_field_length= sort_field->length -
sort_field->suffix_length;
if (sort_field_length < length)
{
diff= 0;
length= sort_field_length;
}
else
diff= sort_field_length - length;
if (sort_field->suffix_length)
{
/* Store length last in result_string */
store_length(to + sort_field_length, length,
sort_field->suffix_length);
}
/* apply cs->sort_order for case-insensitive comparison if needed */
my_strnxfrm(cs,(uchar*)to,length,(const uchar*)res->ptr(),length);
cs->cset->fill(cs, (char *)to+length,diff,fill_char);
}
break;
}
case INT_RESULT:
case TIME_RESULT:
{
longlong UNINIT_VAR(value);
if (sort_field->result_type == INT_RESULT)
value= item->val_int_result();
else
{
MYSQL_TIME buf;
if (item->get_date_result(&buf, TIME_INVALID_DATES))
{
DBUG_ASSERT(maybe_null);
DBUG_ASSERT(item->null_value);
}
else
value= pack_time(&buf);
}
if (maybe_null)
{
*to++=1; /* purecov: inspected */
if (item->null_value)
{
if (maybe_null)
memset(to-1, 0, sort_field->length+1);
else
{
DBUG_PRINT("warning",
("Got null on something that shouldn't be null"));
memset(to, 0, sort_field->length);
}
break;
}
}
to[7]= (uchar) value;
to[6]= (uchar) (value >> 8);
to[5]= (uchar) (value >> 16);
to[4]= (uchar) (value >> 24);
to[3]= (uchar) (value >> 32);
to[2]= (uchar) (value >> 40);
to[1]= (uchar) (value >> 48);
if (item->unsigned_flag) /* Fix sign */
to[0]= (uchar) (value >> 56);
else
to[0]= (uchar) (value >> 56) ^ 128; /* Reverse signbit */
break;
}
case DECIMAL_RESULT:
{
my_decimal dec_buf, *dec_val= item->val_decimal_result(&dec_buf);
if (maybe_null)
{
if (item->null_value)
{
memset(to, 0, sort_field->length+1);
to++;
break;
}
*to++=1;
}
my_decimal2binary(E_DEC_FATAL_ERROR, dec_val, to,
item->max_length - (item->decimals ? 1:0),
item->decimals);
break;
}
case REAL_RESULT:
{
double value= item->val_result();
if (maybe_null)
{
if (item->null_value)
{
memset(to, 0, sort_field->length+1);
to++;
break;
}
*to++=1;
}
change_double_for_sort(value,(uchar*) to);
break;
}
case ROW_RESULT:
default:
// This case should never be choosen
DBUG_ASSERT(0);
break;
}
}
if (sort_field->reverse)
{ /* Revers key */
if (maybe_null && (to[-1]= !to[-1]))
{
to+= sort_field->length; // don't waste the time reversing all 0's
continue;
}
length=sort_field->length;
while (length--)
{
*to = (uchar) (~ *to);
to++;
}
}
else
to+= sort_field->length;
}
if (param->addon_field)
{
/*
Save field values appended to sorted fields.
First null bit indicators are appended then field values follow.
In this implementation we use fixed layout for field values -
the same for all records.
*/
SORT_ADDON_FIELD *addonf= param->addon_field;
uchar *nulls= to;
DBUG_ASSERT(addonf != 0);
memset(nulls, 0, addonf->offset);
to+= addonf->offset;
for ( ; (field= addonf->field) ; addonf++)
{
if (addonf->null_bit && field->is_null())
{
nulls[addonf->null_offset]|= addonf->null_bit;
#ifdef HAVE_valgrind
bzero(to, addonf->length);
#endif
}
else
{
#ifdef HAVE_valgrind
uchar *end= field->pack(to, field->ptr);
uint length= (uint) ((to + addonf->length) - end);
DBUG_ASSERT((int) length >= 0);
if (length)
bzero(end, length);
#else
(void) field->pack(to, field->ptr);
#endif
}
to+= addonf->length;
}
}
else
{
/* Save filepos last */
memcpy((uchar*) to, ref_pos, (size_t) param->ref_length);
}
return;
}
/*
Register fields used by sorting in the sorted table's read set
*/
static void register_used_fields(Sort_param *param)
{
reg1 SORT_FIELD *sort_field;
TABLE *table=param->sort_form;
MY_BITMAP *bitmap= table->read_set;
for (sort_field= param->local_sortorder ;
sort_field != param->end ;
sort_field++)
{
Field *field;
if ((field= sort_field->field))
{
if (field->table == table)
{
if (field->vcol_info)
{
Item *vcol_item= field->vcol_info->expr_item;
vcol_item->walk(&Item::register_field_in_read_map, 1, (uchar *) 0);
}
bitmap_set_bit(bitmap, field->field_index);
}
}
else
{ // Item
sort_field->item->walk(&Item::register_field_in_read_map, 1,
(uchar *) table);
}
}
if (param->addon_field)
{
SORT_ADDON_FIELD *addonf= param->addon_field;
Field *field;
for ( ; (field= addonf->field) ; addonf++)
bitmap_set_bit(bitmap, field->field_index);
}
else
{
/* Save filepos last */
table->prepare_for_position();
}
}
static bool save_index(Sort_param *param, uint count, Filesort_info *table_sort)
{
uint offset,res_length;
uchar *to;
DBUG_ENTER("save_index");
table_sort->sort_buffer(param, count);
res_length= param->res_length;
offset= param->rec_length-res_length;
if (!(to= table_sort->record_pointers=
(uchar*) my_malloc(res_length*count,
MYF(MY_WME | MY_THREAD_SPECIFIC))))
DBUG_RETURN(1); /* purecov: inspected */
uchar **sort_keys= table_sort->get_sort_keys();
for (uchar **end= sort_keys+count ; sort_keys != end ; sort_keys++)
{
memcpy(to, *sort_keys+offset, res_length);
to+= res_length;
}
DBUG_RETURN(0);
}
/**
Test whether priority queue is worth using to get top elements of an
ordered result set. If it is, then allocates buffer for required amount of
records
@param param Sort parameters.
@param filesort_info Filesort information.
@param table Table to sort.
@param num_rows Estimate of number of rows in source record set.
@param memory_available Memory available for sorting.
DESCRIPTION
Given a query like this:
SELECT ... FROM t ORDER BY a1,...,an LIMIT max_rows;
This function tests whether a priority queue should be used to keep
the result. Necessary conditions are:
- estimate that it is actually cheaper than merge-sort
- enough memory to store the <max_rows> records.
If we don't have space for <max_rows> records, but we *do* have
space for <max_rows> keys, we may rewrite 'table' to sort with
references to records instead of additional data.
(again, based on estimates that it will actually be cheaper).
@retval
true - if it's ok to use PQ
false - PQ will be slower than merge-sort, or there is not enough memory.
*/
bool check_if_pq_applicable(Sort_param *param,
Filesort_info *filesort_info,
TABLE *table, ha_rows num_rows,
ulong memory_available)
{
DBUG_ENTER("check_if_pq_applicable");
/*
How much Priority Queue sort is slower than qsort.
Measurements (see unit test) indicate that PQ is roughly 3 times slower.
*/
const double PQ_slowness= 3.0;
if (param->max_rows == HA_POS_ERROR)
{
DBUG_PRINT("info", ("No LIMIT"));
DBUG_RETURN(false);
}
if (param->max_rows + 2 >= UINT_MAX)
{
DBUG_PRINT("info", ("Too large LIMIT"));
DBUG_RETURN(false);
}
ulong num_available_keys=
memory_available / (param->rec_length + sizeof(char*));
// We need 1 extra record in the buffer, when using PQ.
param->max_keys_per_buffer= (uint) param->max_rows + 1;
if (num_rows < num_available_keys)
{
// The whole source set fits into memory.
if (param->max_rows < num_rows/PQ_slowness )
{
filesort_info->alloc_sort_buffer(param->max_keys_per_buffer,
param->rec_length);
DBUG_RETURN(filesort_info->get_sort_keys() != NULL);
}
else
{
// PQ will be slower.
DBUG_RETURN(false);
}
}
// Do we have space for LIMIT rows in memory?
if (param->max_keys_per_buffer < num_available_keys)
{
filesort_info->alloc_sort_buffer(param->max_keys_per_buffer,
param->rec_length);
DBUG_RETURN(filesort_info->get_sort_keys() != NULL);
}
// Try to strip off addon fields.
if (param->addon_field)
{
const ulong row_length=
param->sort_length + param->ref_length + sizeof(char*);
num_available_keys= memory_available / row_length;
// Can we fit all the keys in memory?
if (param->max_keys_per_buffer < num_available_keys)
{
const double sort_merge_cost=
get_merge_many_buffs_cost_fast(num_rows,
num_available_keys,
row_length);
/*
PQ has cost:
(insert + qsort) * log(queue size) / TIME_FOR_COMPARE_ROWID +
cost of file lookup afterwards.
The lookup cost is a bit pessimistic: we take scan_time and assume
that on average we find the row after scanning half of the file.
A better estimate would be lookup cost, but note that we are doing
random lookups here, rather than sequential scan.
*/
const double pq_cpu_cost=
(PQ_slowness * num_rows + param->max_keys_per_buffer) *
log((double) param->max_keys_per_buffer) / TIME_FOR_COMPARE_ROWID;
const double pq_io_cost=
param->max_rows * table->file->scan_time() / 2.0;
const double pq_cost= pq_cpu_cost + pq_io_cost;
if (sort_merge_cost < pq_cost)
DBUG_RETURN(false);
filesort_info->alloc_sort_buffer(param->max_keys_per_buffer,
param->sort_length + param->ref_length);
if (filesort_info->get_sort_keys())
{
// Make attached data to be references instead of fields.
my_free(filesort_info->addon_buf);
my_free(filesort_info->addon_field);
filesort_info->addon_buf= NULL;
filesort_info->addon_field= NULL;
param->addon_field= NULL;
param->addon_length= 0;
param->res_length= param->ref_length;
param->sort_length+= param->ref_length;
param->rec_length= param->sort_length;
DBUG_RETURN(true);
}
}
}
DBUG_RETURN(false);
}
/** Merge buffers to make < MERGEBUFF2 buffers. */
int merge_many_buff(Sort_param *param, uchar *sort_buffer,
BUFFPEK *buffpek, uint *maxbuffer, IO_CACHE *t_file)
{
register uint i;
IO_CACHE t_file2,*from_file,*to_file,*temp;
BUFFPEK *lastbuff;
DBUG_ENTER("merge_many_buff");
if (*maxbuffer < MERGEBUFF2)
DBUG_RETURN(0); /* purecov: inspected */
if (flush_io_cache(t_file) ||
open_cached_file(&t_file2,mysql_tmpdir,TEMP_PREFIX,DISK_BUFFER_SIZE,
MYF(MY_WME)))
DBUG_RETURN(1); /* purecov: inspected */
from_file= t_file ; to_file= &t_file2;
while (*maxbuffer >= MERGEBUFF2)
{
if (reinit_io_cache(from_file,READ_CACHE,0L,0,0))
goto cleanup;
if (reinit_io_cache(to_file,WRITE_CACHE,0L,0,0))
goto cleanup;
lastbuff=buffpek;
for (i=0 ; i <= *maxbuffer-MERGEBUFF*3/2 ; i+=MERGEBUFF)
{
if (merge_buffers(param,from_file,to_file,sort_buffer,lastbuff++,
buffpek+i,buffpek+i+MERGEBUFF-1,0))
goto cleanup;
}
if (merge_buffers(param,from_file,to_file,sort_buffer,lastbuff++,
buffpek+i,buffpek+ *maxbuffer,0))
break; /* purecov: inspected */
if (flush_io_cache(to_file))
break; /* purecov: inspected */
temp=from_file; from_file=to_file; to_file=temp;
setup_io_cache(from_file);
setup_io_cache(to_file);
*maxbuffer= (uint) (lastbuff-buffpek)-1;
}
cleanup:
close_cached_file(to_file); // This holds old result
if (to_file == t_file)
{
*t_file=t_file2; // Copy result file
setup_io_cache(t_file);
}
DBUG_RETURN(*maxbuffer >= MERGEBUFF2); /* Return 1 if interrupted */
} /* merge_many_buff */
/**
Read data to buffer.
@retval
(uint)-1 if something goes wrong
*/
uint read_to_buffer(IO_CACHE *fromfile, BUFFPEK *buffpek,
uint rec_length)
{
register uint count;
uint length;
if ((count=(uint) MY_MIN((ha_rows) buffpek->max_keys,buffpek->count)))
{
if (mysql_file_pread(fromfile->file, (uchar*) buffpek->base,
(length= rec_length*count),
buffpek->file_pos, MYF_RW))
return((uint) -1); /* purecov: inspected */
buffpek->key=buffpek->base;
buffpek->file_pos+= length; /* New filepos */
buffpek->count-= count;
buffpek->mem_count= count;
}
return (count*rec_length);
} /* read_to_buffer */
/**
Put all room used by freed buffer to use in adjacent buffer.
Note, that we can't simply distribute memory evenly between all buffers,
because new areas must not overlap with old ones.
@param[in] queue list of non-empty buffers, without freed buffer
@param[in] reuse empty buffer
@param[in] key_length key length
*/
void reuse_freed_buff(QUEUE *queue, BUFFPEK *reuse, uint key_length)
{
uchar *reuse_end= reuse->base + reuse->max_keys * key_length;
for (uint i= queue_first_element(queue);
i <= queue_last_element(queue);
i++)
{
BUFFPEK *bp= (BUFFPEK *) queue_element(queue, i);
if (bp->base + bp->max_keys * key_length == reuse->base)
{
bp->max_keys+= reuse->max_keys;
return;
}
else if (bp->base == reuse_end)
{
bp->base= reuse->base;
bp->max_keys+= reuse->max_keys;
return;
}
}
DBUG_ASSERT(0);
}
/**
Merge buffers to one buffer.
@param param Sort parameter
@param from_file File with source data (BUFFPEKs point to this file)
@param to_file File to write the sorted result data.
@param sort_buffer Buffer for data to store up to MERGEBUFF2 sort keys.
@param lastbuff OUT Store here BUFFPEK describing data written to to_file
@param Fb First element in source BUFFPEKs array
@param Tb Last element in source BUFFPEKs array
@param flag
@retval
0 OK
@retval
other error
*/
int merge_buffers(Sort_param *param, IO_CACHE *from_file,
IO_CACHE *to_file, uchar *sort_buffer,
BUFFPEK *lastbuff, BUFFPEK *Fb, BUFFPEK *Tb,
int flag)
{
int error;
uint rec_length,res_length,offset;
size_t sort_length;
ulong maxcount;
ha_rows max_rows,org_max_rows;
my_off_t to_start_filepos;
uchar *strpos;
BUFFPEK *buffpek;
QUEUE queue;
qsort2_cmp cmp;
void *first_cmp_arg;
element_count dupl_count= 0;
uchar *src;
uchar *unique_buff= param->unique_buff;
const bool killable= !param->not_killable;
THD* const thd=current_thd;
DBUG_ENTER("merge_buffers");
thd->inc_status_sort_merge_passes();
thd->query_plan_fsort_passes++;
error=0;
rec_length= param->rec_length;
res_length= param->res_length;
sort_length= param->sort_length;
uint dupl_count_ofs= rec_length-sizeof(element_count);
uint min_dupl_count= param->min_dupl_count;
bool check_dupl_count= flag && min_dupl_count;
offset= (rec_length-
(flag && min_dupl_count ? sizeof(dupl_count) : 0)-res_length);
uint wr_len= flag ? res_length : rec_length;
uint wr_offset= flag ? offset : 0;
maxcount= (ulong) (param->max_keys_per_buffer/((uint) (Tb-Fb) +1));
to_start_filepos= my_b_tell(to_file);
strpos= sort_buffer;
org_max_rows=max_rows= param->max_rows;
set_if_bigger(maxcount, 1);
if (unique_buff)
{
cmp= param->compare;
first_cmp_arg= (void *) ¶m->cmp_context;
}
else
{
cmp= get_ptr_compare(sort_length);
first_cmp_arg= (void*) &sort_length;
}
if (init_queue(&queue, (uint) (Tb-Fb)+1, offsetof(BUFFPEK,key), 0,
(queue_compare) cmp, first_cmp_arg, 0, 0))
DBUG_RETURN(1); /* purecov: inspected */
for (buffpek= Fb ; buffpek <= Tb ; buffpek++)
{
buffpek->base= strpos;
buffpek->max_keys= maxcount;
strpos+=
(uint) (error= (int) read_to_buffer(from_file, buffpek, rec_length));
if (error == -1)
goto err; /* purecov: inspected */
buffpek->max_keys= buffpek->mem_count; // If less data in buffers than expected
queue_insert(&queue, (uchar*) buffpek);
}
if (unique_buff)
{
/*
Called by Unique::get()
Copy the first argument to unique_buff for unique removal.
Store it also in 'to_file'.
*/
buffpek= (BUFFPEK*) queue_top(&queue);
memcpy(unique_buff, buffpek->key, rec_length);
if (min_dupl_count)
memcpy(&dupl_count, unique_buff+dupl_count_ofs,
sizeof(dupl_count));
buffpek->key+= rec_length;
if (! --buffpek->mem_count)
{
if (!(error= (int) read_to_buffer(from_file, buffpek,
rec_length)))
{
queue_remove(&queue,0);
reuse_freed_buff(&queue, buffpek, rec_length);
}
else if (error == -1)
goto err; /* purecov: inspected */
}
queue_replace_top(&queue); // Top element has been used
}
else
cmp= 0; // Not unique
while (queue.elements > 1)
{
if (killable && thd->check_killed())
{
error= 1; goto err; /* purecov: inspected */
}
for (;;)
{
buffpek= (BUFFPEK*) queue_top(&queue);
src= buffpek->key;
if (cmp) // Remove duplicates
{
if (!(*cmp)(first_cmp_arg, &unique_buff,
(uchar**) &buffpek->key))
{
if (min_dupl_count)
{
element_count cnt;
memcpy(&cnt, (uchar *) buffpek->key+dupl_count_ofs, sizeof(cnt));
dupl_count+= cnt;
}
goto skip_duplicate;
}
if (min_dupl_count)
{
memcpy(unique_buff+dupl_count_ofs, &dupl_count,
sizeof(dupl_count));
}
src= unique_buff;
}
/*
Do not write into the output file if this is the final merge called
for a Unique object used for intersection and dupl_count is less
than min_dupl_count.
If the Unique object is used to intersect N sets of unique elements
then for any element:
dupl_count >= N <=> the element is occurred in each of these N sets.
*/
if (!check_dupl_count || dupl_count >= min_dupl_count)
{
if (my_b_write(to_file, src+wr_offset, wr_len))
{
error=1; goto err; /* purecov: inspected */
}
}
if (cmp)
{
memcpy(unique_buff, (uchar*) buffpek->key, rec_length);
if (min_dupl_count)
memcpy(&dupl_count, unique_buff+dupl_count_ofs,
sizeof(dupl_count));
}
if (!--max_rows)
{
error= 0; /* purecov: inspected */
goto end; /* purecov: inspected */
}
skip_duplicate:
buffpek->key+= rec_length;
if (! --buffpek->mem_count)
{
if (!(error= (int) read_to_buffer(from_file, buffpek,
rec_length)))
{
(void) queue_remove_top(&queue);
reuse_freed_buff(&queue, buffpek, rec_length);
break; /* One buffer have been removed */
}
else if (error == -1)
goto err; /* purecov: inspected */
}
queue_replace_top(&queue); /* Top element has been replaced */
}
}
buffpek= (BUFFPEK*) queue_top(&queue);
buffpek->base= (uchar*) sort_buffer;
buffpek->max_keys= param->max_keys_per_buffer;
/*
As we know all entries in the buffer are unique, we only have to
check if the first one is the same as the last one we wrote
*/
if (cmp)
{
if (!(*cmp)(first_cmp_arg, &unique_buff, (uchar**) &buffpek->key))
{
if (min_dupl_count)
{
element_count cnt;
memcpy(&cnt, (uchar *) buffpek->key+dupl_count_ofs, sizeof(cnt));
dupl_count+= cnt;
}
buffpek->key+= rec_length;
--buffpek->mem_count;
}
if (min_dupl_count)
memcpy(unique_buff+dupl_count_ofs, &dupl_count,
sizeof(dupl_count));
if (!check_dupl_count || dupl_count >= min_dupl_count)
{
src= unique_buff;
if (my_b_write(to_file, src+wr_offset, wr_len))
{
error=1; goto err; /* purecov: inspected */
}
if (!--max_rows)
{
error= 0;
goto end;
}
}
}
do
{
if ((ha_rows) buffpek->mem_count > max_rows)
{ /* Don't write too many records */
buffpek->mem_count= (uint) max_rows;
buffpek->count= 0; /* Don't read more */
}
max_rows-= buffpek->mem_count;
if (flag == 0)
{
if (my_b_write(to_file, (uchar*) buffpek->key,
(rec_length*buffpek->mem_count)))
{
error= 1; goto err; /* purecov: inspected */
}
}
else
{
register uchar *end;
src= buffpek->key+offset;
for (end= src+buffpek->mem_count*rec_length ;
src != end ;
src+= rec_length)
{
if (check_dupl_count)
{
memcpy((uchar *) &dupl_count, src+dupl_count_ofs, sizeof(dupl_count));
if (dupl_count < min_dupl_count)
continue;
}
if (my_b_write(to_file, src, wr_len))
{
error=1; goto err;
}
}
}
}
while ((error=(int) read_to_buffer(from_file, buffpek, rec_length))
!= -1 && error != 0);
end:
lastbuff->count= MY_MIN(org_max_rows-max_rows, param->max_rows);
lastbuff->file_pos= to_start_filepos;
err:
delete_queue(&queue);
DBUG_RETURN(error);
} /* merge_buffers */
/* Do a merge to output-file (save only positions) */
int merge_index(Sort_param *param, uchar *sort_buffer,
BUFFPEK *buffpek, uint maxbuffer,
IO_CACHE *tempfile, IO_CACHE *outfile)
{
DBUG_ENTER("merge_index");
if (merge_buffers(param,tempfile,outfile,sort_buffer,buffpek,buffpek,
buffpek+maxbuffer,1))
DBUG_RETURN(1); /* purecov: inspected */
DBUG_RETURN(0);
} /* merge_index */
static uint suffix_length(ulong string_length)
{
if (string_length < 256)
return 1;
if (string_length < 256L*256L)
return 2;
if (string_length < 256L*256L*256L)
return 3;
return 4; // Can't sort longer than 4G
}
/**
Calculate length of sort key.
@param thd Thread handler
@param sortorder Order of items to sort
@param s_length Number of items to sort
@param[out] multi_byte_charset Set to 1 if we are using multi-byte charset
(In which case we have to use strxnfrm())
@note
sortorder->length is updated for each sort item.
@n
sortorder->need_strxnfrm is set 1 if we have to use strxnfrm
@return
Total length of sort buffer in bytes
*/
static uint
sortlength(THD *thd, SORT_FIELD *sortorder, uint s_length,
bool *multi_byte_charset)
{
reg2 uint length;
const CHARSET_INFO *cs;
*multi_byte_charset= 0;
length=0;
for (; s_length-- ; sortorder++)
{
sortorder->need_strxnfrm= 0;
sortorder->suffix_length= 0;
if (sortorder->field)
{
cs= sortorder->field->sort_charset();
sortorder->length= sortorder->field->sort_length();
if (use_strnxfrm((cs=sortorder->field->sort_charset())))
{
sortorder->need_strxnfrm= 1;
*multi_byte_charset= 1;
sortorder->length= cs->coll->strnxfrmlen(cs, sortorder->length);
}
if (sortorder->field->maybe_null())
length++; // Place for NULL marker
}
else
{
sortorder->result_type= sortorder->item->cmp_type();
switch (sortorder->result_type) {
case STRING_RESULT:
sortorder->length=sortorder->item->max_length;
set_if_smaller(sortorder->length, thd->variables.max_sort_length);
if (use_strnxfrm((cs=sortorder->item->collation.collation)))
{
sortorder->length= cs->coll->strnxfrmlen(cs, sortorder->length);
sortorder->need_strxnfrm= 1;
*multi_byte_charset= 1;
}
else if (cs == &my_charset_bin)
{
/* Store length last to be able to sort blob/varbinary */
sortorder->suffix_length= suffix_length(sortorder->length);
sortorder->length+= sortorder->suffix_length;
}
break;
case TIME_RESULT:
case INT_RESULT:
sortorder->length=8; // Size of intern longlong
break;
case DECIMAL_RESULT:
sortorder->length=
my_decimal_get_binary_size(sortorder->item->max_length -
(sortorder->item->decimals ? 1 : 0),
sortorder->item->decimals);
break;
case REAL_RESULT:
sortorder->length=sizeof(double);
break;
case ROW_RESULT:
default:
// This case should never be choosen
DBUG_ASSERT(0);
break;
}
if (sortorder->item->maybe_null)
length++; // Place for NULL marker
}
set_if_smaller(sortorder->length, thd->variables.max_sort_length);
length+=sortorder->length;
}
sortorder->field= (Field*) 0; // end marker
DBUG_PRINT("info",("sort_length: %d",length));
return length;
}
/**
Get descriptors of fields appended to sorted fields and
calculate its total length.
The function first finds out what fields are used in the result set.
Then it calculates the length of the buffer to store the values of
these fields together with the value of sort values.
If the calculated length is not greater than max_length_for_sort_data
the function allocates memory for an array of descriptors containing
layouts for the values of the non-sorted fields in the buffer and
fills them.
@param thd Current thread
@param ptabfield Array of references to the table fields
@param sortlength Total length of sorted fields
@param[out] plength Total length of appended fields
@note
The null bits for the appended values are supposed to be put together
and stored the buffer just ahead of the value of the first field.
@return
Pointer to the layout descriptors for the appended fields, if any
@retval
NULL if we do not store field values with sort data.
*/
static SORT_ADDON_FIELD *
get_addon_fields(ulong max_length_for_sort_data,
Field **ptabfield, uint sortlength, uint *plength)
{
Field **pfield;
Field *field;
SORT_ADDON_FIELD *addonf;
uint length= 0;
uint fields= 0;
uint null_fields= 0;
MY_BITMAP *read_set= (*ptabfield)->table->read_set;
/*
If there is a reference to a field in the query add it
to the the set of appended fields.
Note for future refinement:
This this a too strong condition.
Actually we need only the fields referred in the
result set. And for some of them it makes sense to use
the values directly from sorted fields.
But beware the case when item->cmp_type() != item->result_type()
*/
*plength= 0;
for (pfield= ptabfield; (field= *pfield) ; pfield++)
{
if (!bitmap_is_set(read_set, field->field_index))
continue;
if (field->flags & BLOB_FLAG)
return 0;
length+= field->max_packed_col_length(field->pack_length());
if (field->maybe_null())
null_fields++;
fields++;
}
if (!fields)
return 0;
length+= (null_fields+7)/8;
if (length+sortlength > max_length_for_sort_data ||
!(addonf= (SORT_ADDON_FIELD *) my_malloc(sizeof(SORT_ADDON_FIELD)*
(fields+1),
MYF(MY_WME |
MY_THREAD_SPECIFIC))))
return 0;
*plength= length;
length= (null_fields+7)/8;
null_fields= 0;
for (pfield= ptabfield; (field= *pfield) ; pfield++)
{
if (!bitmap_is_set(read_set, field->field_index))
continue;
addonf->field= field;
addonf->offset= length;
if (field->maybe_null())
{
addonf->null_offset= null_fields/8;
addonf->null_bit= 1<<(null_fields & 7);
null_fields++;
}
else
{
addonf->null_offset= 0;
addonf->null_bit= 0;
}
addonf->length= field->max_packed_col_length(field->pack_length());
length+= addonf->length;
addonf++;
}
addonf->field= 0; // Put end marker
DBUG_PRINT("info",("addon_length: %d",length));
return (addonf-fields);
}
/**
Copy (unpack) values appended to sorted fields from a buffer back to
their regular positions specified by the Field::ptr pointers.
@param addon_field Array of descriptors for appended fields
@param buff Buffer which to unpack the value from
@note
The function is supposed to be used only as a callback function
when getting field values for the sorted result set.
@return
void.
*/
static void
unpack_addon_fields(struct st_sort_addon_field *addon_field, uchar *buff,
uchar *buff_end)
{
Field *field;
SORT_ADDON_FIELD *addonf= addon_field;
for ( ; (field= addonf->field) ; addonf++)
{
if (addonf->null_bit && (addonf->null_bit & buff[addonf->null_offset]))
{
field->set_null();
continue;
}
field->set_notnull();
field->unpack(field->ptr, buff + addonf->offset, buff_end, 0);
}
}
/*
** functions to change a double or float to a sortable string
** The following should work for IEEE
*/
#define DBL_EXP_DIG (sizeof(double)*8-DBL_MANT_DIG)
void change_double_for_sort(double nr,uchar *to)
{
uchar *tmp=(uchar*) to;
if (nr == 0.0)
{ /* Change to zero string */
tmp[0]=(uchar) 128;
memset(tmp+1, 0, sizeof(nr)-1);
}
else
{
#ifdef WORDS_BIGENDIAN
memcpy(tmp, &nr, sizeof(nr));
#else
{
uchar *ptr= (uchar*) &nr;
#if defined(__FLOAT_WORD_ORDER) && (__FLOAT_WORD_ORDER == __BIG_ENDIAN)
tmp[0]= ptr[3]; tmp[1]=ptr[2]; tmp[2]= ptr[1]; tmp[3]=ptr[0];
tmp[4]= ptr[7]; tmp[5]=ptr[6]; tmp[6]= ptr[5]; tmp[7]=ptr[4];
#else
tmp[0]= ptr[7]; tmp[1]=ptr[6]; tmp[2]= ptr[5]; tmp[3]=ptr[4];
tmp[4]= ptr[3]; tmp[5]=ptr[2]; tmp[6]= ptr[1]; tmp[7]=ptr[0];
#endif
}
#endif
if (tmp[0] & 128) /* Negative */
{ /* make complement */
uint i;
for (i=0 ; i < sizeof(nr); i++)
tmp[i]=tmp[i] ^ (uchar) 255;
}
else
{ /* Set high and move exponent one up */
ushort exp_part=(((ushort) tmp[0] << 8) | (ushort) tmp[1] |
(ushort) 32768);
exp_part+= (ushort) 1 << (16-1-DBL_EXP_DIG);
tmp[0]= (uchar) (exp_part >> 8);
tmp[1]= (uchar) exp_part;
}
}
}
|