File: sqlite.py

package info (click to toggle)
python-ase 3.22.1-3
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 14,344 kB
  • sloc: python: 126,379; xml: 946; makefile: 111; javascript: 47
file content (928 lines) | stat: -rw-r--r-- 33,495 bytes parent folder | download
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
"""SQLite3 backend.

Versions:

1) Added 3 more columns.
2) Changed "user" to "username".
3) Now adding keys to keyword table and added an "information" table containing
   a version number.
4) Got rid of keywords.
5) Add fmax, smax, mass, volume, charge
6) Use REAL for magmom and drop possibility for non-collinear spin
7) Volume can be None
8) Added name='metadata' row to "information" table
9) Row data is now stored in binary format.
"""

import json
import numbers
import os
import sqlite3
import sys
from contextlib import contextmanager

import numpy as np

import ase.io.jsonio
from ase.data import atomic_numbers
from ase.calculators.calculator import all_properties
from ase.db.row import AtomsRow
from ase.db.core import (Database, ops, now, lock, invop, parse_selection,
                         object_to_bytes, bytes_to_object)
from ase.parallel import parallel_function

VERSION = 9

init_statements = [
    """CREATE TABLE systems (
    id INTEGER PRIMARY KEY AUTOINCREMENT,  -- ID's, timestamps and user name
    unique_id TEXT UNIQUE,
    ctime REAL,
    mtime REAL,
    username TEXT,
    numbers BLOB,  -- stuff that defines an Atoms object
    positions BLOB,
    cell BLOB,
    pbc INTEGER,
    initial_magmoms BLOB,
    initial_charges BLOB,
    masses BLOB,
    tags BLOB,
    momenta BLOB,
    constraints TEXT,  -- constraints and calculator
    calculator TEXT,
    calculator_parameters TEXT,
    energy REAL,  -- calculated properties
    free_energy REAL,
    forces BLOB,
    stress BLOB,
    dipole BLOB,
    magmoms BLOB,
    magmom REAL,
    charges BLOB,
    key_value_pairs TEXT,  -- key-value pairs and data as json
    data BLOB,
    natoms INTEGER,  -- stuff for making queries faster
    fmax REAL,
    smax REAL,
    volume REAL,
    mass REAL,
    charge REAL)""",

    """CREATE TABLE species (
    Z INTEGER,
    n INTEGER,
    id INTEGER,
    FOREIGN KEY (id) REFERENCES systems(id))""",

    """CREATE TABLE keys (
    key TEXT,
    id INTEGER,
    FOREIGN KEY (id) REFERENCES systems(id))""",

    """CREATE TABLE text_key_values (
    key TEXT,
    value TEXT,
    id INTEGER,
    FOREIGN KEY (id) REFERENCES systems(id))""",

    """CREATE TABLE number_key_values (
    key TEXT,
    value REAL,
    id INTEGER,
    FOREIGN KEY (id) REFERENCES systems(id))""",

    """CREATE TABLE information (
    name TEXT,
    value TEXT)""",

    "INSERT INTO information VALUES ('version', '{}')".format(VERSION)]

index_statements = [
    'CREATE INDEX unique_id_index ON systems(unique_id)',
    'CREATE INDEX ctime_index ON systems(ctime)',
    'CREATE INDEX username_index ON systems(username)',
    'CREATE INDEX calculator_index ON systems(calculator)',
    'CREATE INDEX species_index ON species(Z)',
    'CREATE INDEX key_index ON keys(key)',
    'CREATE INDEX text_index ON text_key_values(key)',
    'CREATE INDEX number_index ON number_key_values(key)']

all_tables = ['systems', 'species', 'keys',
              'text_key_values', 'number_key_values']


def float_if_not_none(x):
    """Convert numpy.float64 to float - old db-interfaces need that."""
    if x is not None:
        return float(x)


class SQLite3Database(Database):
    type = 'db'
    initialized = False
    _allow_reading_old_format = False
    default = 'NULL'  # used for autoincrement id
    connection = None
    version = None
    columnnames = [line.split()[0].lstrip()
                   for line in init_statements[0].splitlines()[1:]]

    def encode(self, obj, binary=False):
        if binary:
            return object_to_bytes(obj)
        return ase.io.jsonio.encode(obj)

    def decode(self, txt, lazy=False):
        if lazy:
            return txt
        if isinstance(txt, str):
            return ase.io.jsonio.decode(txt)
        return bytes_to_object(txt)

    def blob(self, array):
        """Convert array to blob/buffer object."""

        if array is None:
            return None
        if len(array) == 0:
            array = np.zeros(0)
        if array.dtype == np.int64:
            array = array.astype(np.int32)
        if not np.little_endian:
            array = array.byteswap()
        return memoryview(np.ascontiguousarray(array))

    def deblob(self, buf, dtype=float, shape=None):
        """Convert blob/buffer object to ndarray of correct dtype and shape.

        (without creating an extra view)."""
        if buf is None:
            return None
        if len(buf) == 0:
            array = np.zeros(0, dtype)
        else:
            array = np.frombuffer(buf, dtype)
            if not np.little_endian:
                array = array.byteswap()
        if shape is not None:
            array.shape = shape
        return array

    def _connect(self):
        return sqlite3.connect(self.filename, timeout=20)

    def __enter__(self):
        assert self.connection is None
        self.change_count = 0
        self.connection = self._connect()
        return self

    def __exit__(self, exc_type, exc_value, tb):
        if exc_type is None:
            self.connection.commit()
        else:
            self.connection.rollback()
        self.connection.close()
        self.connection = None

    @contextmanager
    def managed_connection(self, commit_frequency=5000):
        try:
            con = self.connection or self._connect()
            self._initialize(con)
            yield con
        except ValueError as exc:
            if self.connection is None:
                con.close()
            raise exc
        else:
            if self.connection is None:
                con.commit()
                con.close()
            else:
                self.change_count += 1
                if self.change_count % commit_frequency == 0:
                    con.commit()

    def _initialize(self, con):
        if self.initialized:
            return

        self._metadata = {}

        cur = con.execute(
            'SELECT COUNT(*) FROM sqlite_master WHERE name="systems"')

        if cur.fetchone()[0] == 0:
            for statement in init_statements:
                con.execute(statement)
            if self.create_indices:
                for statement in index_statements:
                    con.execute(statement)
            con.commit()
            self.version = VERSION
        else:
            cur = con.execute(
                'SELECT COUNT(*) FROM sqlite_master WHERE name="user_index"')
            if cur.fetchone()[0] == 1:
                # Old version with "user" instead of "username" column
                self.version = 1
            else:
                try:
                    cur = con.execute(
                        'SELECT value FROM information WHERE name="version"')
                except sqlite3.OperationalError:
                    self.version = 2
                else:
                    self.version = int(cur.fetchone()[0])

                cur = con.execute(
                    'SELECT value FROM information WHERE name="metadata"')
                results = cur.fetchall()
                if results:
                    self._metadata = json.loads(results[0][0])

        if self.version > VERSION:
            raise IOError('Can not read new ase.db format '
                          '(version {}).  Please update to latest ASE.'
                          .format(self.version))
        if self.version < 5 and not self._allow_reading_old_format:
            raise IOError('Please convert to new format. ' +
                          'Use: python -m ase.db.convert ' + self.filename)

        self.initialized = True

    def _write(self, atoms, key_value_pairs, data, id):
        ext_tables = key_value_pairs.pop("external_tables", {})
        Database._write(self, atoms, key_value_pairs, data)

        mtime = now()

        encode = self.encode
        blob = self.blob

        if not isinstance(atoms, AtomsRow):
            row = AtomsRow(atoms)
            row.ctime = mtime
            row.user = os.getenv('USER')
        else:
            row = atoms
            # Extract the external tables from AtomsRow
            names = self._get_external_table_names()
            for name in names:
                new_table = row.get(name, {})
                if new_table:
                    ext_tables[name] = new_table

        if not id and not key_value_pairs and not ext_tables:
            key_value_pairs = row.key_value_pairs

        for k, v in ext_tables.items():
            dtype = self._guess_type(v)
            self._create_table_if_not_exists(k, dtype)

        constraints = row._constraints
        if constraints:
            if isinstance(constraints, list):
                constraints = encode(constraints)
        else:
            constraints = None

        values = (row.unique_id,
                  row.ctime,
                  mtime,
                  row.user,
                  blob(row.numbers),
                  blob(row.positions),
                  blob(row.cell),
                  int(np.dot(row.pbc, [1, 2, 4])),
                  blob(row.get('initial_magmoms')),
                  blob(row.get('initial_charges')),
                  blob(row.get('masses')),
                  blob(row.get('tags')),
                  blob(row.get('momenta')),
                  constraints)

        if 'calculator' in row:
            values += (row.calculator, encode(row.calculator_parameters))
        else:
            values += (None, None)

        if not data:
            data = row._data

        with self.managed_connection() as con:
            if not isinstance(data, (str, bytes)):
                data = encode(data, binary=self.version >= 9)

            values += (row.get('energy'),
                       row.get('free_energy'),
                       blob(row.get('forces')),
                       blob(row.get('stress')),
                       blob(row.get('dipole')),
                       blob(row.get('magmoms')),
                       row.get('magmom'),
                       blob(row.get('charges')),
                       encode(key_value_pairs),
                       data,
                       len(row.numbers),
                       float_if_not_none(row.get('fmax')),
                       float_if_not_none(row.get('smax')),
                       float_if_not_none(row.get('volume')),
                       float(row.mass),
                       float(row.charge))

            cur = con.cursor()
            if id is None:
                q = self.default + ', ' + ', '.join('?' * len(values))
                cur.execute('INSERT INTO systems VALUES ({})'.format(q),
                            values)
                id = self.get_last_id(cur)
            else:
                self._delete(cur, [id], ['keys', 'text_key_values',
                                         'number_key_values', 'species'])
                q = ', '.join(name + '=?' for name in self.columnnames[1:])
                cur.execute('UPDATE systems SET {} WHERE id=?'.format(q),
                            values + (id,))

            count = row.count_atoms()
            if count:
                species = [(atomic_numbers[symbol], n, id)
                           for symbol, n in count.items()]
                cur.executemany('INSERT INTO species VALUES (?, ?, ?)',
                                species)

            text_key_values = []
            number_key_values = []
            for key, value in key_value_pairs.items():
                if isinstance(value, (numbers.Real, np.bool_)):
                    number_key_values.append([key, float(value), id])
                else:
                    assert isinstance(value, str)
                    text_key_values.append([key, value, id])

            cur.executemany('INSERT INTO text_key_values VALUES (?, ?, ?)',
                            text_key_values)
            cur.executemany('INSERT INTO number_key_values VALUES (?, ?, ?)',
                            number_key_values)
            cur.executemany('INSERT INTO keys VALUES (?, ?)',
                            [(key, id) for key in key_value_pairs])

            # Insert entries in the valid tables
            for tabname in ext_tables.keys():
                entries = ext_tables[tabname]
                entries['id'] = id
                self._insert_in_external_table(
                    cur, name=tabname, entries=ext_tables[tabname])

        return id

    def _update(self, id, key_value_pairs, data=None):
        """Update key_value_pairs and data for a single row """
        encode = self.encode
        ext_tables = key_value_pairs.pop('external_tables', {})

        for k, v in ext_tables.items():
            dtype = self._guess_type(v)
            self._create_table_if_not_exists(k, dtype)

        mtime = now()
        with self.managed_connection() as con:
            cur = con.cursor()
            cur.execute(
                'UPDATE systems SET mtime=?, key_value_pairs=? WHERE id=?',
                (mtime, encode(key_value_pairs), id))
            if data:
                if not isinstance(data, (str, bytes)):
                    data = encode(data, binary=self.version >= 9)
                cur.execute('UPDATE systems set data=? where id=?', (data, id))

            self._delete(cur, [id], ['keys', 'text_key_values',
                                     'number_key_values'])

            text_key_values = []
            number_key_values = []
            for key, value in key_value_pairs.items():
                if isinstance(value, (numbers.Real, np.bool_)):
                    number_key_values.append([key, float(value), id])
                else:
                    assert isinstance(value, str)
                    text_key_values.append([key, value, id])

            cur.executemany('INSERT INTO text_key_values VALUES (?, ?, ?)',
                            text_key_values)
            cur.executemany('INSERT INTO number_key_values VALUES (?, ?, ?)',
                            number_key_values)
            cur.executemany('INSERT INTO keys VALUES (?, ?)',
                            [(key, id) for key in key_value_pairs])

            # Insert entries in the valid tables
            for tabname in ext_tables.keys():
                entries = ext_tables[tabname]
                entries['id'] = id
                self._insert_in_external_table(
                    cur, name=tabname, entries=ext_tables[tabname])

        return id

    def get_last_id(self, cur):
        cur.execute('SELECT seq FROM sqlite_sequence WHERE name="systems"')
        result = cur.fetchone()
        if result is not None:
            id = result[0]
            return id
        else:
            return 0

    def _get_row(self, id):
        with self.managed_connection() as con:
            cur = con.cursor()
            if id is None:
                cur.execute('SELECT COUNT(*) FROM systems')
                assert cur.fetchone()[0] == 1
                cur.execute('SELECT * FROM systems')
            else:
                cur.execute('SELECT * FROM systems WHERE id=?', (id,))
            values = cur.fetchone()

        return self._convert_tuple_to_row(values)

    def _convert_tuple_to_row(self, values):
        deblob = self.deblob
        decode = self.decode

        values = self._old2new(values)
        dct = {'id': values[0],
               'unique_id': values[1],
               'ctime': values[2],
               'mtime': values[3],
               'user': values[4],
               'numbers': deblob(values[5], np.int32),
               'positions': deblob(values[6], shape=(-1, 3)),
               'cell': deblob(values[7], shape=(3, 3))}

        if values[8] is not None:
            dct['pbc'] = (values[8] & np.array([1, 2, 4])).astype(bool)
        if values[9] is not None:
            dct['initial_magmoms'] = deblob(values[9])
        if values[10] is not None:
            dct['initial_charges'] = deblob(values[10])
        if values[11] is not None:
            dct['masses'] = deblob(values[11])
        if values[12] is not None:
            dct['tags'] = deblob(values[12], np.int32)
        if values[13] is not None:
            dct['momenta'] = deblob(values[13], shape=(-1, 3))
        if values[14] is not None:
            dct['constraints'] = values[14]
        if values[15] is not None:
            dct['calculator'] = values[15]
        if values[16] is not None:
            dct['calculator_parameters'] = decode(values[16])
        if values[17] is not None:
            dct['energy'] = values[17]
        if values[18] is not None:
            dct['free_energy'] = values[18]
        if values[19] is not None:
            dct['forces'] = deblob(values[19], shape=(-1, 3))
        if values[20] is not None:
            dct['stress'] = deblob(values[20])
        if values[21] is not None:
            dct['dipole'] = deblob(values[21])
        if values[22] is not None:
            dct['magmoms'] = deblob(values[22])
        if values[23] is not None:
            dct['magmom'] = values[23]
        if values[24] is not None:
            dct['charges'] = deblob(values[24])
        if values[25] != '{}':
            dct['key_value_pairs'] = decode(values[25])
        if len(values) >= 27 and values[26] != 'null':
            dct['data'] = decode(values[26], lazy=True)

        # Now we need to update with info from the external tables
        external_tab = self._get_external_table_names()
        tables = {}
        for tab in external_tab:
            row = self._read_external_table(tab, dct["id"])
            tables[tab] = row

        dct.update(tables)
        return AtomsRow(dct)

    def _old2new(self, values):
        if self.type == 'postgresql':
            assert self.version >= 8, 'Your db-version is too old!'
        assert self.version >= 4, 'Your db-file is too old!'
        if self.version < 5:
            pass  # should be ok for reading by convert.py script
        if self.version < 6:
            m = values[23]
            if m is not None and not isinstance(m, float):
                magmom = float(self.deblob(m, shape=()))
                values = values[:23] + (magmom,) + values[24:]
        return values

    def create_select_statement(self, keys, cmps,
                                sort=None, order=None, sort_table=None,
                                what='systems.*'):
        tables = ['systems']
        where = []
        args = []
        for key in keys:
            if key == 'forces':
                where.append('systems.fmax IS NOT NULL')
            elif key == 'strain':
                where.append('systems.smax IS NOT NULL')
            elif key in ['energy', 'fmax', 'smax',
                         'constraints', 'calculator']:
                where.append('systems.{} IS NOT NULL'.format(key))
            else:
                if '-' not in key:
                    q = 'systems.id in (select id from keys where key=?)'
                else:
                    key = key.replace('-', '')
                    q = 'systems.id not in (select id from keys where key=?)'
                where.append(q)
                args.append(key)

        # Special handling of "H=0" and "H<2" type of selections:
        bad = {}
        for key, op, value in cmps:
            if isinstance(key, int):
                bad[key] = bad.get(key, True) and ops[op](0, value)

        for key, op, value in cmps:
            if key in ['id', 'energy', 'magmom', 'ctime', 'user',
                       'calculator', 'natoms', 'pbc', 'unique_id',
                       'fmax', 'smax', 'volume', 'mass', 'charge']:
                if key == 'user':
                    key = 'username'
                elif key == 'pbc':
                    assert op in ['=', '!=']
                    value = int(np.dot([x == 'T' for x in value], [1, 2, 4]))
                elif key == 'magmom':
                    assert self.version >= 6, 'Update your db-file'
                where.append('systems.{}{}?'.format(key, op))
                args.append(value)
            elif isinstance(key, int):
                if self.type == 'postgresql':
                    where.append(
                        'cardinality(array_positions(' +
                        'numbers::int[], ?)){}?'.format(op))
                    args += [key, value]
                else:
                    if bad[key]:
                        where.append(
                            'systems.id not in (select id from species ' +
                            'where Z=? and n{}?)'.format(invop[op]))
                        args += [key, value]
                    else:
                        where.append('systems.id in (select id from species ' +
                                     'where Z=? and n{}?)'.format(op))
                        args += [key, value]

            elif self.type == 'postgresql':
                jsonop = '->'
                if isinstance(value, str):
                    jsonop = '->>'
                elif isinstance(value, bool):
                    jsonop = '->>'
                    value = str(value).lower()
                where.append("systems.key_value_pairs {} '{}'{}?"
                             .format(jsonop, key, op))
                args.append(str(value))

            elif isinstance(value, str):
                where.append('systems.id in (select id from text_key_values ' +
                             'where key=? and value{}?)'.format(op))
                args += [key, value]
            else:
                where.append(
                    'systems.id in (select id from number_key_values ' +
                    'where key=? and value{}?)'.format(op))
                args += [key, float(value)]

        if sort:
            if sort_table != 'systems':
                tables.append('{} AS sort_table'.format(sort_table))
                where.append('systems.id=sort_table.id AND '
                             'sort_table.key=?')
                args.append(sort)
                sort_table = 'sort_table'
                sort = 'value'

        sql = 'SELECT {} FROM\n  '.format(what) + ', '.join(tables)
        if where:
            sql += '\n  WHERE\n  ' + ' AND\n  '.join(where)
        if sort:
            # XXX use "?" instead of "{}"
            sql += '\nORDER BY {0}.{1} IS NULL, {0}.{1} {2}'.format(
                sort_table, sort, order)

        return sql, args

    def _select(self, keys, cmps, explain=False, verbosity=0,
                limit=None, offset=0, sort=None, include_data=True,
                columns='all'):

        values = np.array([None for i in range(27)])
        values[25] = '{}'
        values[26] = 'null'

        if columns == 'all':
            columnindex = list(range(26))
        else:
            columnindex = [c for c in range(0, 26)
                           if self.columnnames[c] in columns]
        if include_data:
            columnindex.append(26)

        if sort:
            if sort[0] == '-':
                order = 'DESC'
                sort = sort[1:]
            else:
                order = 'ASC'
            if sort in ['id', 'energy', 'username', 'calculator',
                        'ctime', 'mtime', 'magmom', 'pbc',
                        'fmax', 'smax', 'volume', 'mass', 'charge', 'natoms']:
                sort_table = 'systems'
            else:
                for dct in self._select(keys + [sort], cmps=[], limit=1,
                                        include_data=False,
                                        columns=['key_value_pairs']):
                    if isinstance(dct['key_value_pairs'][sort], str):
                        sort_table = 'text_key_values'
                    else:
                        sort_table = 'number_key_values'
                    break
                else:
                    # No rows.  Just pick a table:
                    sort_table = 'number_key_values'

        else:
            order = None
            sort_table = None

        what = ', '.join('systems.' + name
                         for name in
                         np.array(self.columnnames)[np.array(columnindex)])

        sql, args = self.create_select_statement(keys, cmps, sort, order,
                                                 sort_table, what)

        if explain:
            sql = 'EXPLAIN QUERY PLAN ' + sql

        if limit:
            sql += '\nLIMIT {0}'.format(limit)

        if offset:
            sql += self.get_offset_string(offset, limit=limit)

        if verbosity == 2:
            print(sql, args)

        with self.managed_connection() as con:
            cur = con.cursor()
            cur.execute(sql, args)
            if explain:
                for row in cur.fetchall():
                    yield {'explain': row}
            else:
                n = 0
                for shortvalues in cur.fetchall():
                    values[columnindex] = shortvalues
                    yield self._convert_tuple_to_row(tuple(values))
                    n += 1

                if sort and sort_table != 'systems':
                    # Yield rows without sort key last:
                    if limit is not None:
                        if n == limit:
                            return
                        limit -= n
                    for row in self._select(keys + ['-' + sort], cmps,
                                            limit=limit, offset=offset,
                                            include_data=include_data,
                                            columns=columns):
                        yield row

    def get_offset_string(self, offset, limit=None):
        sql = ''
        if not limit:
            # In sqlite you cannot have offset without limit, so we
            # set it to -1 meaning no limit
            sql += '\nLIMIT -1'
        sql += '\nOFFSET {0}'.format(offset)
        return sql

    @parallel_function
    def count(self, selection=None, **kwargs):
        keys, cmps = parse_selection(selection, **kwargs)
        sql, args = self.create_select_statement(keys, cmps, what='COUNT(*)')

        with self.managed_connection() as con:
            cur = con.cursor()
            cur.execute(sql, args)
            return cur.fetchone()[0]

    def analyse(self):
        with self.managed_connection() as con:
            con.execute('ANALYZE')

    @parallel_function
    @lock
    def delete(self, ids):
        if len(ids) == 0:
            return
        table_names = self._get_external_table_names() + all_tables[::-1]
        with self.managed_connection() as con:
            self._delete(con.cursor(), ids,
                         tables=table_names)
        self.vacuum()

    def _delete(self, cur, ids, tables=None):
        tables = tables or all_tables[::-1]
        for table in tables:
            cur.execute('DELETE FROM {} WHERE id in ({});'.
                        format(table, ', '.join([str(id) for id in ids])))

    def vacuum(self):
        if not self.type == 'db':
            return

        with self.managed_connection() as con:
            con.commit()
            con.cursor().execute("VACUUM")

    @property
    def metadata(self):
        if self._metadata is None:
            self._initialize(self._connect())
        return self._metadata.copy()

    @metadata.setter
    def metadata(self, dct):
        self._metadata = dct
        md = json.dumps(dct)
        with self.managed_connection() as con:
            cur = con.cursor()
            cur.execute(
                "SELECT COUNT(*) FROM information WHERE name='metadata'")

            if cur.fetchone()[0]:
                cur.execute(
                    "UPDATE information SET value=? WHERE name='metadata'",
                    [md])
            else:
                cur.execute('INSERT INTO information VALUES (?, ?)',
                            ('metadata', md))

    def _get_external_table_names(self, db_con=None):
        """Return a list with the external table names."""
        sql = "SELECT value FROM information WHERE name='external_table_name'"
        with self.managed_connection() as con:
            cur = con.cursor()
            cur.execute(sql)
            ext_tab_names = [x[0] for x in cur.fetchall()]
        return ext_tab_names

    def _external_table_exists(self, name):
        """Return True if an external table name exists."""
        return name in self._get_external_table_names()

    def _create_table_if_not_exists(self, name, dtype):
        """Create a new table if it does not exits.

        Arguments
        ==========
        name: str
            Name of the new table
        dtype: str
            Datatype of the value field (typically REAL, INTEGER, TEXT etc.)
        """

        taken_names = set(all_tables + all_properties + self.columnnames)
        if name in taken_names:
            raise ValueError("External table can not be any of {}"
                             "".format(taken_names))

        if self._external_table_exists(name):
            return

        sql = "CREATE TABLE IF NOT EXISTS {} ".format(name)
        sql += "(key TEXT, value {}, id INTEGER, ".format(dtype)
        sql += "FOREIGN KEY (id) REFERENCES systems(id))"
        sql2 = "INSERT INTO information VALUES (?, ?)"
        with self.managed_connection() as con:
            cur = con.cursor()
            cur.execute(sql)
            # Insert an entry saying that there is a new external table
            # present and an entry with the datatype
            cur.execute(sql2, ("external_table_name", name))
            cur.execute(sql2, (name + "_dtype", dtype))

    def delete_external_table(self, name):
        """Delete an external table."""
        if not self._external_table_exists(name):
            return

        with self.managed_connection() as con:
            cur = con.cursor()

            sql = "DROP TABLE {}".format(name)
            cur.execute(sql)

            sql = "DELETE FROM information WHERE value=?"
            cur.execute(sql, (name,))
            sql = "DELETE FROM information WHERE name=?"
            cur.execute(sql, (name + "_dtype",))

    def _convert_to_recognized_types(self, value):
        """Convert Numpy types to python types."""
        if np.issubdtype(type(value), np.integer):
            return int(value)
        elif np.issubdtype(type(value), np.floating):
            return float(value)
        return value

    def _insert_in_external_table(self, cursor, name=None, entries=None):
        """Insert into external table"""
        if name is None or entries is None:
            # There is nothing to do
            return

        id = entries.pop("id")
        dtype = self._guess_type(entries)
        expected_dtype = self._get_value_type_of_table(cursor, name)
        if dtype != expected_dtype:
            raise ValueError("The provided data type for table {} "
                             "is {}, while it is initialized to "
                             "be of type {}"
                             "".format(name, dtype, expected_dtype))

        # First we check if entries already exists
        cursor.execute("SELECT key FROM {} WHERE id=?".format(name), (id,))
        updates = []
        for item in cursor.fetchall():
            value = entries.pop(item[0], None)
            if value is not None:
                updates.append(
                    (value, id, self._convert_to_recognized_types(item[0])))

        # Update entry if key and ID already exists
        sql = "UPDATE {} SET value=? WHERE id=? AND key=?".format(name)
        cursor.executemany(sql, updates)

        # Insert the ones that does not already exist
        inserts = [(k, self._convert_to_recognized_types(v), id)
                   for k, v in entries.items()]
        sql = "INSERT INTO {} VALUES (?, ?, ?)".format(name)
        cursor.executemany(sql, inserts)

    def _guess_type(self, entries):
        """Guess the type based on the first entry."""
        values = [v for _, v in entries.items()]

        # Check if all datatypes are the same
        all_types = [type(v) for v in values]
        if any([t != all_types[0] for t in all_types]):
            typenames = [t.__name__ for t in all_types]
            raise ValueError("Inconsistent datatypes in the table. "
                             "given types: {}".format(typenames))

        val = values[0]
        if isinstance(val, int) or np.issubdtype(type(val), np.integer):
            return "INTEGER"
        if isinstance(val, float) or np.issubdtype(type(val), np.floating):
            return "REAL"
        if isinstance(val, str):
            return "TEXT"
        raise ValueError("Unknown datatype!")

    def _get_value_type_of_table(self, cursor, tab_name):
        """Return the expected value name."""
        sql = "SELECT value FROM information WHERE name=?"
        cursor.execute(sql, (tab_name + "_dtype",))
        return cursor.fetchone()[0]

    def _read_external_table(self, name, id):
        """Read row from external table."""

        with self.managed_connection() as con:
            cur = con.cursor()
            cur.execute("SELECT * FROM {} WHERE id=?".format(name), (id,))
            items = cur.fetchall()
            dictionary = dict([(item[0], item[1]) for item in items])

        return dictionary


if __name__ == '__main__':
    from ase.db import connect
    con = connect(sys.argv[1])
    con._initialize(con._connect())
    print('Version:', con.version)