1
|
# Database access
|
2
|
|
3
|
import copy
|
4
|
import re
|
5
|
import warnings
|
6
|
|
7
|
import exc
|
8
|
import dicts
|
9
|
import iters
|
10
|
import lists
|
11
|
from Proxy import Proxy
|
12
|
import rand
|
13
|
import sql_gen
|
14
|
import strings
|
15
|
import util
|
16
|
|
17
|
##### Exceptions
|
18
|
|
19
|
def get_cur_query(cur, input_query=None):
|
20
|
raw_query = None
|
21
|
if hasattr(cur, 'query'): raw_query = cur.query
|
22
|
elif hasattr(cur, '_last_executed'): raw_query = cur._last_executed
|
23
|
|
24
|
if raw_query != None: return raw_query
|
25
|
else: return '[input] '+strings.ustr(input_query)
|
26
|
|
27
|
def _add_cursor_info(e, *args, **kw_args):
|
28
|
'''For params, see get_cur_query()'''
|
29
|
exc.add_msg(e, 'query: '+strings.ustr(get_cur_query(*args, **kw_args)))
|
30
|
|
31
|
class DbException(exc.ExceptionWithCause):
|
32
|
def __init__(self, msg, cause=None, cur=None):
|
33
|
exc.ExceptionWithCause.__init__(self, msg, cause, cause_newline=True)
|
34
|
if cur != None: _add_cursor_info(self, cur)
|
35
|
|
36
|
class ExceptionWithName(DbException):
|
37
|
def __init__(self, name, cause=None):
|
38
|
DbException.__init__(self, 'for name: '+strings.as_tt(str(name)), cause)
|
39
|
self.name = name
|
40
|
|
41
|
class ExceptionWithValue(DbException):
|
42
|
def __init__(self, value, cause=None):
|
43
|
DbException.__init__(self, 'for value: '+strings.as_tt(repr(value)),
|
44
|
cause)
|
45
|
self.value = value
|
46
|
|
47
|
class ExceptionWithNameType(DbException):
|
48
|
def __init__(self, type_, name, cause=None):
|
49
|
DbException.__init__(self, 'for type: '+strings.as_tt(str(type_))
|
50
|
+'; name: '+strings.as_tt(name), cause)
|
51
|
self.type = type_
|
52
|
self.name = name
|
53
|
|
54
|
class ConstraintException(DbException):
|
55
|
def __init__(self, name, cols, cause=None):
|
56
|
DbException.__init__(self, 'Violated '+strings.as_tt(name)
|
57
|
+' constraint on columns: '+strings.as_tt(', '.join(cols)), cause)
|
58
|
self.name = name
|
59
|
self.cols = cols
|
60
|
|
61
|
class MissingCastException(DbException):
|
62
|
def __init__(self, type_, col, cause=None):
|
63
|
DbException.__init__(self, 'Missing cast to type '+strings.as_tt(type_)
|
64
|
+' on column: '+strings.as_tt(col), cause)
|
65
|
self.type = type_
|
66
|
self.col = col
|
67
|
|
68
|
class NameException(DbException): pass
|
69
|
|
70
|
class DuplicateKeyException(ConstraintException): pass
|
71
|
|
72
|
class NullValueException(ConstraintException): pass
|
73
|
|
74
|
class InvalidValueException(ExceptionWithValue): pass
|
75
|
|
76
|
class DuplicateException(ExceptionWithNameType): pass
|
77
|
|
78
|
class EmptyRowException(DbException): pass
|
79
|
|
80
|
##### Warnings
|
81
|
|
82
|
class DbWarning(UserWarning): pass
|
83
|
|
84
|
##### Result retrieval
|
85
|
|
86
|
def col_names(cur): return (col[0] for col in cur.description)
|
87
|
|
88
|
def rows(cur): return iter(lambda: cur.fetchone(), None)
|
89
|
|
90
|
def consume_rows(cur):
|
91
|
'''Used to fetch all rows so result will be cached'''
|
92
|
iters.consume_iter(rows(cur))
|
93
|
|
94
|
def next_row(cur): return rows(cur).next()
|
95
|
|
96
|
def row(cur):
|
97
|
row_ = next_row(cur)
|
98
|
consume_rows(cur)
|
99
|
return row_
|
100
|
|
101
|
def next_value(cur): return next_row(cur)[0]
|
102
|
|
103
|
def value(cur): return row(cur)[0]
|
104
|
|
105
|
def values(cur): return iters.func_iter(lambda: next_value(cur))
|
106
|
|
107
|
def value_or_none(cur):
|
108
|
try: return value(cur)
|
109
|
except StopIteration: return None
|
110
|
|
111
|
##### Escaping
|
112
|
|
113
|
def esc_name_by_module(module, name):
|
114
|
if module == 'psycopg2' or module == None: quote = '"'
|
115
|
elif module == 'MySQLdb': quote = '`'
|
116
|
else: raise NotImplementedError("Can't escape name for "+module+' database')
|
117
|
return sql_gen.esc_name(name, quote)
|
118
|
|
119
|
def esc_name_by_engine(engine, name, **kw_args):
|
120
|
return esc_name_by_module(db_engines[engine][0], name, **kw_args)
|
121
|
|
122
|
def esc_name(db, name, **kw_args):
|
123
|
return esc_name_by_module(util.root_module(db.db), name, **kw_args)
|
124
|
|
125
|
def qual_name(db, schema, table):
|
126
|
def esc_name_(name): return esc_name(db, name)
|
127
|
table = esc_name_(table)
|
128
|
if schema != None: return esc_name_(schema)+'.'+table
|
129
|
else: return table
|
130
|
|
131
|
##### Database connections
|
132
|
|
133
|
db_config_names = ['engine', 'host', 'user', 'password', 'database', 'schemas']
|
134
|
|
135
|
db_engines = {
|
136
|
'MySQL': ('MySQLdb', {'password': 'passwd', 'database': 'db'}),
|
137
|
'PostgreSQL': ('psycopg2', {}),
|
138
|
}
|
139
|
|
140
|
DatabaseErrors_set = set([DbException])
|
141
|
DatabaseErrors = tuple(DatabaseErrors_set)
|
142
|
|
143
|
def _add_module(module):
|
144
|
DatabaseErrors_set.add(module.DatabaseError)
|
145
|
global DatabaseErrors
|
146
|
DatabaseErrors = tuple(DatabaseErrors_set)
|
147
|
|
148
|
def db_config_str(db_config):
|
149
|
return db_config['engine']+' database '+db_config['database']
|
150
|
|
151
|
log_debug_none = lambda msg, level=2: None
|
152
|
|
153
|
class DbConn:
|
154
|
def __init__(self, db_config, autocommit=True, caching=True,
|
155
|
log_debug=log_debug_none, debug_temp=False):
|
156
|
'''
|
157
|
@param debug_temp Whether temporary objects should instead be permanent.
|
158
|
This assists in debugging the internal objects used by the program.
|
159
|
'''
|
160
|
self.db_config = db_config
|
161
|
self.autocommit = autocommit
|
162
|
self.caching = caching
|
163
|
self.log_debug = log_debug
|
164
|
self.debug = log_debug != log_debug_none
|
165
|
self.debug_temp = debug_temp
|
166
|
self.autoanalyze = False
|
167
|
|
168
|
self._savepoint = 0
|
169
|
self._reset()
|
170
|
|
171
|
def __getattr__(self, name):
|
172
|
if name == '__dict__': raise Exception('getting __dict__')
|
173
|
if name == 'db': return self._db()
|
174
|
else: raise AttributeError()
|
175
|
|
176
|
def __getstate__(self):
|
177
|
state = copy.copy(self.__dict__) # shallow copy
|
178
|
state['log_debug'] = None # don't pickle the debug callback
|
179
|
state['_DbConn__db'] = None # don't pickle the connection
|
180
|
return state
|
181
|
|
182
|
def clear_cache(self): self.query_results = {}
|
183
|
|
184
|
def _reset(self):
|
185
|
self.clear_cache()
|
186
|
assert self._savepoint == 0
|
187
|
self._notices_seen = set()
|
188
|
self.__db = None
|
189
|
|
190
|
def connected(self): return self.__db != None
|
191
|
|
192
|
def close(self):
|
193
|
if not self.connected(): return
|
194
|
|
195
|
# Record that the automatic transaction is now closed
|
196
|
self._savepoint -= 1
|
197
|
|
198
|
self.db.close()
|
199
|
self._reset()
|
200
|
|
201
|
def reconnect(self):
|
202
|
# Do not do this in test mode as it would roll back everything
|
203
|
if self.autocommit: self.close()
|
204
|
# Connection will be reopened automatically on first query
|
205
|
|
206
|
def _db(self):
|
207
|
if self.__db == None:
|
208
|
# Process db_config
|
209
|
db_config = self.db_config.copy() # don't modify input!
|
210
|
schemas = db_config.pop('schemas', None)
|
211
|
module_name, mappings = db_engines[db_config.pop('engine')]
|
212
|
module = __import__(module_name)
|
213
|
_add_module(module)
|
214
|
for orig, new in mappings.iteritems():
|
215
|
try: util.rename_key(db_config, orig, new)
|
216
|
except KeyError: pass
|
217
|
|
218
|
# Connect
|
219
|
self.__db = module.connect(**db_config)
|
220
|
|
221
|
# Configure connection
|
222
|
if hasattr(self.db, 'set_isolation_level'):
|
223
|
import psycopg2.extensions
|
224
|
self.db.set_isolation_level(
|
225
|
psycopg2.extensions.ISOLATION_LEVEL_READ_COMMITTED)
|
226
|
if schemas != None:
|
227
|
search_path = [self.esc_name(s) for s in schemas.split(',')]
|
228
|
search_path.append(value(run_query(self, 'SHOW search_path',
|
229
|
log_level=4)))
|
230
|
run_query(self, 'SET search_path TO '+(','.join(search_path)),
|
231
|
log_level=3)
|
232
|
|
233
|
# Record that a transaction is already open
|
234
|
self._savepoint += 1
|
235
|
|
236
|
return self.__db
|
237
|
|
238
|
class DbCursor(Proxy):
|
239
|
def __init__(self, outer):
|
240
|
Proxy.__init__(self, outer.db.cursor())
|
241
|
self.outer = outer
|
242
|
self.query_results = outer.query_results
|
243
|
self.query_lookup = None
|
244
|
self.result = []
|
245
|
|
246
|
def execute(self, query):
|
247
|
self._is_insert = query.startswith('INSERT')
|
248
|
self.query_lookup = query
|
249
|
try:
|
250
|
try:
|
251
|
cur = self.inner.execute(query)
|
252
|
self.outer.do_autocommit()
|
253
|
finally: self.query = get_cur_query(self.inner, query)
|
254
|
except Exception, e:
|
255
|
_add_cursor_info(e, self, query)
|
256
|
self.result = e # cache the exception as the result
|
257
|
self._cache_result()
|
258
|
raise
|
259
|
|
260
|
# Always cache certain queries
|
261
|
if query.startswith('CREATE') or query.startswith('ALTER'):
|
262
|
# structural changes
|
263
|
# Rest of query must be unique in the face of name collisions,
|
264
|
# so don't cache ADD COLUMN unless it has distinguishing comment
|
265
|
if query.find('ADD COLUMN') < 0 or query.endswith('*/'):
|
266
|
self._cache_result()
|
267
|
elif self.rowcount == 0 and query.startswith('SELECT'): # empty
|
268
|
consume_rows(self) # fetch all rows so result will be cached
|
269
|
|
270
|
return cur
|
271
|
|
272
|
def fetchone(self):
|
273
|
row = self.inner.fetchone()
|
274
|
if row != None: self.result.append(row)
|
275
|
# otherwise, fetched all rows
|
276
|
else: self._cache_result()
|
277
|
return row
|
278
|
|
279
|
def _cache_result(self):
|
280
|
# For inserts that return a result set, don't cache result set since
|
281
|
# inserts are not idempotent. Other non-SELECT queries don't have
|
282
|
# their result set read, so only exceptions will be cached (an
|
283
|
# invalid query will always be invalid).
|
284
|
if self.query_results != None and (not self._is_insert
|
285
|
or isinstance(self.result, Exception)):
|
286
|
|
287
|
assert self.query_lookup != None
|
288
|
self.query_results[self.query_lookup] = self.CacheCursor(
|
289
|
util.dict_subset(dicts.AttrsDictView(self),
|
290
|
['query', 'result', 'rowcount', 'description']))
|
291
|
|
292
|
class CacheCursor:
|
293
|
def __init__(self, cached_result): self.__dict__ = cached_result
|
294
|
|
295
|
def execute(self, *args, **kw_args):
|
296
|
if isinstance(self.result, Exception): raise self.result
|
297
|
# otherwise, result is a rows list
|
298
|
self.iter = iter(self.result)
|
299
|
|
300
|
def fetchone(self):
|
301
|
try: return self.iter.next()
|
302
|
except StopIteration: return None
|
303
|
|
304
|
def esc_value(self, value):
|
305
|
try: str_ = self.mogrify('%s', [value])
|
306
|
except NotImplementedError, e:
|
307
|
module = util.root_module(self.db)
|
308
|
if module == 'MySQLdb':
|
309
|
import _mysql
|
310
|
str_ = _mysql.escape_string(value)
|
311
|
else: raise e
|
312
|
return strings.to_unicode(str_)
|
313
|
|
314
|
def esc_name(self, name): return esc_name(self, name) # calls global func
|
315
|
|
316
|
def std_code(self, str_):
|
317
|
'''Standardizes SQL code.
|
318
|
* Ensures that string literals are prefixed by `E`
|
319
|
'''
|
320
|
if str_.startswith("'"): str_ = 'E'+str_
|
321
|
return str_
|
322
|
|
323
|
def can_mogrify(self):
|
324
|
module = util.root_module(self.db)
|
325
|
return module == 'psycopg2'
|
326
|
|
327
|
def mogrify(self, query, params=None):
|
328
|
if self.can_mogrify(): return self.db.cursor().mogrify(query, params)
|
329
|
else: raise NotImplementedError("Can't mogrify query")
|
330
|
|
331
|
def print_notices(self):
|
332
|
if hasattr(self.db, 'notices'):
|
333
|
for msg in self.db.notices:
|
334
|
if msg not in self._notices_seen:
|
335
|
self._notices_seen.add(msg)
|
336
|
self.log_debug(msg, level=2)
|
337
|
|
338
|
def run_query(self, query, cacheable=False, log_level=2,
|
339
|
debug_msg_ref=None):
|
340
|
'''
|
341
|
@param log_ignore_excs The log_level will be increased by 2 if the query
|
342
|
throws one of these exceptions.
|
343
|
@param debug_msg_ref If specified, the log message will be returned in
|
344
|
this instead of being output. This allows you to filter log messages
|
345
|
depending on the result of the query.
|
346
|
'''
|
347
|
assert query != None
|
348
|
|
349
|
if not self.caching: cacheable = False
|
350
|
used_cache = False
|
351
|
|
352
|
def log_msg(query):
|
353
|
if used_cache: cache_status = 'cache hit'
|
354
|
elif cacheable: cache_status = 'cache miss'
|
355
|
else: cache_status = 'non-cacheable'
|
356
|
return 'DB query: '+cache_status+':\n'+strings.as_code(query, 'SQL')
|
357
|
|
358
|
try:
|
359
|
# Get cursor
|
360
|
if cacheable:
|
361
|
try:
|
362
|
cur = self.query_results[query]
|
363
|
used_cache = True
|
364
|
except KeyError: cur = self.DbCursor(self)
|
365
|
else: cur = self.db.cursor()
|
366
|
|
367
|
# Log query
|
368
|
if self.debug and debug_msg_ref == None: # log before running
|
369
|
self.log_debug(log_msg(query), log_level)
|
370
|
|
371
|
# Run query
|
372
|
cur.execute(query)
|
373
|
finally:
|
374
|
self.print_notices()
|
375
|
if self.debug and debug_msg_ref != None: # return after running
|
376
|
debug_msg_ref[0] = log_msg(str(get_cur_query(cur, query)))
|
377
|
|
378
|
return cur
|
379
|
|
380
|
def is_cached(self, query): return query in self.query_results
|
381
|
|
382
|
def with_autocommit(self, func):
|
383
|
import psycopg2.extensions
|
384
|
|
385
|
prev_isolation_level = self.db.isolation_level
|
386
|
self.db.set_isolation_level(
|
387
|
psycopg2.extensions.ISOLATION_LEVEL_AUTOCOMMIT)
|
388
|
try: return func()
|
389
|
finally: self.db.set_isolation_level(prev_isolation_level)
|
390
|
|
391
|
def with_savepoint(self, func):
|
392
|
top = self._savepoint == 0
|
393
|
savepoint = 'level_'+str(self._savepoint)
|
394
|
|
395
|
if top: query = 'START TRANSACTION ISOLATION LEVEL READ COMMITTED'
|
396
|
else: query = 'SAVEPOINT '+savepoint
|
397
|
self.run_query(query, log_level=4)
|
398
|
|
399
|
self._savepoint += 1
|
400
|
try:
|
401
|
return func()
|
402
|
if top: self.run_query('COMMIT', log_level=4)
|
403
|
except:
|
404
|
if top: query = 'ROLLBACK'
|
405
|
else: query = 'ROLLBACK TO SAVEPOINT '+savepoint
|
406
|
self.run_query(query, log_level=4)
|
407
|
|
408
|
raise
|
409
|
finally:
|
410
|
# Always release savepoint, because after ROLLBACK TO SAVEPOINT,
|
411
|
# "The savepoint remains valid and can be rolled back to again"
|
412
|
# (http://www.postgresql.org/docs/8.3/static/sql-rollback-to.html).
|
413
|
if not top:
|
414
|
self.run_query('RELEASE SAVEPOINT '+savepoint, log_level=4)
|
415
|
|
416
|
self._savepoint -= 1
|
417
|
assert self._savepoint >= 0
|
418
|
|
419
|
self.do_autocommit() # OK to do this after ROLLBACK TO SAVEPOINT
|
420
|
|
421
|
def do_autocommit(self):
|
422
|
'''Autocommits if outside savepoint'''
|
423
|
assert self._savepoint >= 1
|
424
|
if self.autocommit and self._savepoint == 1:
|
425
|
self.log_debug('Autocommitting', level=4)
|
426
|
self.db.commit()
|
427
|
|
428
|
def col_info(self, col):
|
429
|
table = sql_gen.Table('columns', 'information_schema')
|
430
|
type_ = sql_gen.Coalesce(sql_gen.Nullif(sql_gen.Col('data_type'),
|
431
|
'USER-DEFINED'), sql_gen.Col('udt_name'))
|
432
|
cols = [type_, 'column_default',
|
433
|
sql_gen.Cast('boolean', sql_gen.Col('is_nullable'))]
|
434
|
|
435
|
conds = [('table_name', col.table.name), ('column_name', col.name)]
|
436
|
schema = col.table.schema
|
437
|
if schema != None: conds.append(('table_schema', schema))
|
438
|
|
439
|
type_, default, nullable = row(select(self, table, cols, conds,
|
440
|
order_by='table_schema', limit=1, cacheable=False, log_level=4))
|
441
|
# TODO: order_by search_path schema order
|
442
|
default = sql_gen.as_Code(default, self)
|
443
|
|
444
|
return sql_gen.TypedCol(col.name, type_, default, nullable)
|
445
|
|
446
|
def TempFunction(self, name):
|
447
|
if self.debug_temp: schema = None
|
448
|
else: schema = 'pg_temp'
|
449
|
return sql_gen.Function(name, schema)
|
450
|
|
451
|
connect = DbConn
|
452
|
|
453
|
##### Recoverable querying
|
454
|
|
455
|
def with_savepoint(db, func): return db.with_savepoint(func)
|
456
|
|
457
|
def run_query(db, query, recover=None, cacheable=False, log_level=2,
|
458
|
log_ignore_excs=None, **kw_args):
|
459
|
'''For params, see DbConn.run_query()'''
|
460
|
if recover == None: recover = False
|
461
|
if log_ignore_excs == None: log_ignore_excs = ()
|
462
|
log_ignore_excs = tuple(log_ignore_excs)
|
463
|
|
464
|
debug_msg_ref = None # usually, db.run_query() logs query before running it
|
465
|
# But if filtering with log_ignore_excs, wait until after exception parsing
|
466
|
if log_ignore_excs != () or not db.can_mogrify(): debug_msg_ref = [None]
|
467
|
|
468
|
try:
|
469
|
try:
|
470
|
def run(): return db.run_query(query, cacheable, log_level,
|
471
|
debug_msg_ref, **kw_args)
|
472
|
if recover and not db.is_cached(query):
|
473
|
return with_savepoint(db, run)
|
474
|
else: return run() # don't need savepoint if cached
|
475
|
except Exception, e:
|
476
|
msg = strings.ustr(e.args[0])
|
477
|
|
478
|
match = re.match(r'^duplicate key value violates unique constraint '
|
479
|
r'"((_?[^\W_]+(?=[._]))?.+?)"', msg)
|
480
|
if match:
|
481
|
constraint, table = match.groups()
|
482
|
cols = []
|
483
|
if recover: # need auto-rollback to run index_cols()
|
484
|
try: cols = index_cols(db, table, constraint)
|
485
|
except NotImplementedError: pass
|
486
|
raise DuplicateKeyException(constraint, cols, e)
|
487
|
|
488
|
match = re.match(r'^null value in column "(.+?)" violates not-null'
|
489
|
r' constraint', msg)
|
490
|
if match: raise NullValueException('NOT NULL', [match.group(1)], e)
|
491
|
|
492
|
match = re.match(r'^(?:invalid input (?:syntax|value)\b.*?'
|
493
|
r'|.+? field value out of range): "(.+?)"', msg)
|
494
|
if match:
|
495
|
value, = match.groups()
|
496
|
raise InvalidValueException(strings.to_unicode(value), e)
|
497
|
|
498
|
match = re.match(r'^column "(.+?)" is of type (.+?) but expression '
|
499
|
r'is of type', msg)
|
500
|
if match:
|
501
|
col, type_ = match.groups()
|
502
|
raise MissingCastException(type_, col, e)
|
503
|
|
504
|
match = re.match(r'^(\S+) "(.+?)".*? already exists', msg)
|
505
|
if match:
|
506
|
type_, name = match.groups()
|
507
|
raise DuplicateException(type_, name, e)
|
508
|
|
509
|
raise # no specific exception raised
|
510
|
except log_ignore_excs:
|
511
|
log_level += 2
|
512
|
raise
|
513
|
finally:
|
514
|
if debug_msg_ref != None and debug_msg_ref[0] != None:
|
515
|
db.log_debug(debug_msg_ref[0], log_level)
|
516
|
|
517
|
##### Basic queries
|
518
|
|
519
|
def next_version(name):
|
520
|
version = 1 # first existing name was version 0
|
521
|
match = re.match(r'^(.*)#(\d+)$', name)
|
522
|
if match:
|
523
|
name, version = match.groups()
|
524
|
version = int(version)+1
|
525
|
return sql_gen.concat(name, '#'+str(version))
|
526
|
|
527
|
def lock_table(db, table, mode):
|
528
|
table = sql_gen.as_Table(table)
|
529
|
run_query(db, 'LOCK TABLE '+table.to_str(db)+' IN '+mode+' MODE')
|
530
|
|
531
|
def run_query_into(db, query, into=None, add_indexes_=False, **kw_args):
|
532
|
'''Outputs a query to a temp table.
|
533
|
For params, see run_query().
|
534
|
'''
|
535
|
if into == None: return run_query(db, query, **kw_args)
|
536
|
|
537
|
assert isinstance(into, sql_gen.Table)
|
538
|
|
539
|
into.is_temp = True
|
540
|
# "temporary tables cannot specify a schema name", so remove schema
|
541
|
into.schema = None
|
542
|
|
543
|
kw_args['recover'] = True
|
544
|
kw_args.setdefault('log_ignore_excs', (DuplicateException,))
|
545
|
|
546
|
temp = not db.debug_temp # tables are permanent in debug_temp mode
|
547
|
|
548
|
# Create table
|
549
|
while True:
|
550
|
create_query = 'CREATE'
|
551
|
if temp: create_query += ' TEMP'
|
552
|
create_query += ' TABLE '+into.to_str(db)+' AS\n'+query
|
553
|
|
554
|
try:
|
555
|
cur = run_query(db, create_query, **kw_args)
|
556
|
# CREATE TABLE AS sets rowcount to # rows in query
|
557
|
break
|
558
|
except DuplicateException, e:
|
559
|
into.name = next_version(into.name)
|
560
|
# try again with next version of name
|
561
|
|
562
|
if add_indexes_: add_indexes(db, into)
|
563
|
|
564
|
# According to the PostgreSQL doc, "The autovacuum daemon cannot access and
|
565
|
# therefore cannot vacuum or analyze temporary tables. [...] if a temporary
|
566
|
# table is going to be used in complex queries, it is wise to run ANALYZE on
|
567
|
# the temporary table after it is populated."
|
568
|
# (http://www.postgresql.org/docs/9.1/static/sql-createtable.html)
|
569
|
# If into is not a temp table, ANALYZE is useful but not required.
|
570
|
analyze(db, into)
|
571
|
|
572
|
return cur
|
573
|
|
574
|
order_by_pkey = object() # tells mk_select() to order by the pkey
|
575
|
|
576
|
distinct_on_all = object() # tells mk_select() to SELECT DISTINCT ON all columns
|
577
|
|
578
|
def mk_select(db, tables, fields=None, conds=None, distinct_on=[], limit=None,
|
579
|
start=None, order_by=order_by_pkey, default_table=None):
|
580
|
'''
|
581
|
@param tables The single table to select from, or a list of tables to join
|
582
|
together, with tables after the first being sql_gen.Join objects
|
583
|
@param fields Use None to select all fields in the table
|
584
|
@param conds WHERE conditions: [(compare_left_side, compare_right_side),...]
|
585
|
* container can be any iterable type
|
586
|
* compare_left_side: sql_gen.Code|str (for col name)
|
587
|
* compare_right_side: sql_gen.ValueCond|literal value
|
588
|
@param distinct_on The columns to SELECT DISTINCT ON, or distinct_on_all to
|
589
|
use all columns
|
590
|
@return query
|
591
|
'''
|
592
|
# Parse tables param
|
593
|
tables = lists.mk_seq(tables)
|
594
|
tables = list(tables) # don't modify input! (list() copies input)
|
595
|
table0 = sql_gen.as_Table(tables.pop(0)) # first table is separate
|
596
|
|
597
|
# Parse other params
|
598
|
if conds == None: conds = []
|
599
|
elif dicts.is_dict(conds): conds = conds.items()
|
600
|
conds = list(conds) # don't modify input! (list() copies input)
|
601
|
assert limit == None or isinstance(limit, (int, long))
|
602
|
assert start == None or isinstance(start, (int, long))
|
603
|
if order_by is order_by_pkey:
|
604
|
if distinct_on != []: order_by = None
|
605
|
else: order_by = pkey(db, table0, recover=True)
|
606
|
|
607
|
query = 'SELECT'
|
608
|
|
609
|
def parse_col(col): return sql_gen.as_Col(col, default_table).to_str(db)
|
610
|
|
611
|
# DISTINCT ON columns
|
612
|
if distinct_on != []:
|
613
|
query += '\nDISTINCT'
|
614
|
if distinct_on is not distinct_on_all:
|
615
|
query += ' ON ('+(', '.join(map(parse_col, distinct_on)))+')'
|
616
|
|
617
|
# Columns
|
618
|
if fields == None:
|
619
|
if query.find('\n') >= 0: whitespace = '\n'
|
620
|
else: whitespace = ' '
|
621
|
query += whitespace+'*'
|
622
|
else:
|
623
|
assert fields != []
|
624
|
query += '\n'+('\n, '.join(map(parse_col, fields)))
|
625
|
|
626
|
# Main table
|
627
|
query += '\nFROM '+table0.to_str(db)
|
628
|
|
629
|
# Add joins
|
630
|
left_table = table0
|
631
|
for join_ in tables:
|
632
|
table = join_.table
|
633
|
|
634
|
# Parse special values
|
635
|
if join_.type_ is sql_gen.filter_out: # filter no match
|
636
|
conds.append((sql_gen.Col(table_not_null_col(db, table), table),
|
637
|
sql_gen.CompareCond(None, '~=')))
|
638
|
|
639
|
query += '\n'+join_.to_str(db, left_table)
|
640
|
|
641
|
left_table = table
|
642
|
|
643
|
missing = True
|
644
|
if conds != []:
|
645
|
if len(conds) == 1: whitespace = ' '
|
646
|
else: whitespace = '\n'
|
647
|
query += '\n'+sql_gen.combine_conds([sql_gen.ColValueCond(l, r)
|
648
|
.to_str(db) for l, r in conds], 'WHERE')
|
649
|
missing = False
|
650
|
if order_by != None:
|
651
|
query += '\nORDER BY '+sql_gen.as_Col(order_by, table0).to_str(db)
|
652
|
if limit != None: query += '\nLIMIT '+str(limit); missing = False
|
653
|
if start != None:
|
654
|
if start != 0: query += '\nOFFSET '+str(start)
|
655
|
missing = False
|
656
|
if missing: warnings.warn(DbWarning(
|
657
|
'SELECT statement missing a WHERE, LIMIT, or OFFSET clause: '+query))
|
658
|
|
659
|
return query
|
660
|
|
661
|
def select(db, *args, **kw_args):
|
662
|
'''For params, see mk_select() and run_query()'''
|
663
|
recover = kw_args.pop('recover', None)
|
664
|
cacheable = kw_args.pop('cacheable', True)
|
665
|
log_level = kw_args.pop('log_level', 2)
|
666
|
|
667
|
return run_query(db, mk_select(db, *args, **kw_args), recover, cacheable,
|
668
|
log_level=log_level)
|
669
|
|
670
|
def mk_insert_select(db, table, cols=None, select_query=None, returning=None,
|
671
|
embeddable=False, ignore=False):
|
672
|
'''
|
673
|
@param returning str|None An inserted column (such as pkey) to return
|
674
|
@param embeddable Whether the query should be embeddable as a nested SELECT.
|
675
|
Warning: If you set this and cacheable=True when the query is run, the
|
676
|
query will be fully cached, not just if it raises an exception.
|
677
|
@param ignore Whether to ignore duplicate keys.
|
678
|
'''
|
679
|
table = sql_gen.remove_table_rename(sql_gen.as_Table(table))
|
680
|
if cols == []: cols = None # no cols (all defaults) = unknown col names
|
681
|
if cols != None: cols = [sql_gen.to_name_only_col(c, table) for c in cols]
|
682
|
if select_query == None: select_query = 'DEFAULT VALUES'
|
683
|
if returning != None: returning = sql_gen.as_Col(returning, table)
|
684
|
|
685
|
first_line = 'INSERT INTO '+table.to_str(db)
|
686
|
|
687
|
def mk_insert(select_query):
|
688
|
query = first_line
|
689
|
if cols != None:
|
690
|
query += '\n('+(', '.join((c.to_str(db) for c in cols)))+')'
|
691
|
query += '\n'+select_query
|
692
|
|
693
|
if returning != None:
|
694
|
returning_name_col = sql_gen.to_name_only_col(returning)
|
695
|
query += '\nRETURNING '+returning_name_col.to_str(db)
|
696
|
|
697
|
return query
|
698
|
|
699
|
return_type = 'unknown'
|
700
|
if returning != None: return_type = returning.to_str(db)+'%TYPE'
|
701
|
|
702
|
lang = 'sql'
|
703
|
if ignore:
|
704
|
# Always return something to set the correct rowcount
|
705
|
if returning == None: returning = sql_gen.NamedCol('NULL', None)
|
706
|
|
707
|
embeddable = True # must use function
|
708
|
lang = 'plpgsql'
|
709
|
|
710
|
if cols == None:
|
711
|
row = [sql_gen.Col(sql_gen.all_cols, 'row')]
|
712
|
row_vars = [sql_gen.Table('row')]
|
713
|
else:
|
714
|
row_vars = row = [sql_gen.Col(c.name, 'row') for c in cols]
|
715
|
|
716
|
query = '''\
|
717
|
DECLARE
|
718
|
row '''+table.to_str(db)+'''%ROWTYPE;
|
719
|
BEGIN
|
720
|
/* Need an EXCEPTION block for each individual row because "When an error is
|
721
|
caught by an EXCEPTION clause, [...] all changes to persistent database
|
722
|
state within the block are rolled back."
|
723
|
This is unfortunate because "A block containing an EXCEPTION clause is
|
724
|
significantly more expensive to enter and exit than a block without one."
|
725
|
(http://www.postgresql.org/docs/8.3/static/plpgsql-control-structures.html\
|
726
|
#PLPGSQL-ERROR-TRAPPING)
|
727
|
*/
|
728
|
FOR '''+(', '.join((v.to_str(db) for v in row_vars)))+''' IN
|
729
|
'''+select_query+'''
|
730
|
LOOP
|
731
|
BEGIN
|
732
|
RETURN QUERY
|
733
|
'''+mk_insert(sql_gen.Values(row).to_str(db))+'''
|
734
|
;
|
735
|
EXCEPTION
|
736
|
WHEN unique_violation THEN NULL; -- continue to next row
|
737
|
END;
|
738
|
END LOOP;
|
739
|
END;\
|
740
|
'''
|
741
|
else: query = mk_insert(select_query)
|
742
|
|
743
|
if embeddable:
|
744
|
# Create function
|
745
|
function_name = sql_gen.clean_name(first_line)
|
746
|
while True:
|
747
|
try:
|
748
|
function = db.TempFunction(function_name)
|
749
|
|
750
|
function_query = '''\
|
751
|
CREATE FUNCTION '''+function.to_str(db)+'''()
|
752
|
RETURNS SETOF '''+return_type+'''
|
753
|
LANGUAGE '''+lang+'''
|
754
|
AS $$
|
755
|
'''+query+'''
|
756
|
$$;
|
757
|
'''
|
758
|
run_query(db, function_query, recover=True, cacheable=True,
|
759
|
log_ignore_excs=(DuplicateException,))
|
760
|
break # this version was successful
|
761
|
except DuplicateException, e:
|
762
|
function_name = next_version(function_name)
|
763
|
# try again with next version of name
|
764
|
|
765
|
# Return query that uses function
|
766
|
cols = None
|
767
|
if returning != None: cols = [returning]
|
768
|
func_table = sql_gen.NamedTable('f', sql_gen.FunctionCall(function),
|
769
|
cols) # AS clause requires function alias
|
770
|
return mk_select(db, func_table, start=0, order_by=None)
|
771
|
|
772
|
return query
|
773
|
|
774
|
def insert_select(db, table, *args, **kw_args):
|
775
|
'''For params, see mk_insert_select() and run_query_into()
|
776
|
@param into sql_gen.Table with suggested name of temp table to put RETURNING
|
777
|
values in
|
778
|
'''
|
779
|
returning = kw_args.get('returning', None)
|
780
|
ignore = kw_args.get('ignore', False)
|
781
|
|
782
|
into = kw_args.pop('into', None)
|
783
|
if into != None: kw_args['embeddable'] = True
|
784
|
recover = kw_args.pop('recover', None)
|
785
|
if ignore: recover = True
|
786
|
cacheable = kw_args.pop('cacheable', True)
|
787
|
log_level = kw_args.pop('log_level', 2)
|
788
|
|
789
|
rowcount_only = ignore and returning == None # keep NULL rows on server
|
790
|
if rowcount_only: into = sql_gen.Table('rowcount')
|
791
|
|
792
|
cur = run_query_into(db, mk_insert_select(db, table, *args, **kw_args),
|
793
|
into, recover=recover, cacheable=cacheable, log_level=log_level)
|
794
|
if rowcount_only: empty_temp(db, into)
|
795
|
autoanalyze(db, table)
|
796
|
return cur
|
797
|
|
798
|
default = sql_gen.default # tells insert() to use the default value for a column
|
799
|
|
800
|
def insert(db, table, row, *args, **kw_args):
|
801
|
'''For params, see insert_select()'''
|
802
|
if lists.is_seq(row): cols = None
|
803
|
else:
|
804
|
cols = row.keys()
|
805
|
row = row.values()
|
806
|
row = list(row) # ensure that "== []" works
|
807
|
|
808
|
if row == []: query = None
|
809
|
else: query = sql_gen.Values(row).to_str(db)
|
810
|
|
811
|
return insert_select(db, table, cols, query, *args, **kw_args)
|
812
|
|
813
|
def mk_update(db, table, changes=None, cond=None, in_place=False):
|
814
|
'''
|
815
|
@param changes [(col, new_value),...]
|
816
|
* container can be any iterable type
|
817
|
* col: sql_gen.Code|str (for col name)
|
818
|
* new_value: sql_gen.Code|literal value
|
819
|
@param cond sql_gen.Code WHERE condition. e.g. use sql_gen.*Cond objects.
|
820
|
@param in_place If set, locks the table and updates rows in place.
|
821
|
This avoids creating dead rows in PostgreSQL.
|
822
|
* cond must be None
|
823
|
@return str query
|
824
|
'''
|
825
|
table = sql_gen.as_Table(table)
|
826
|
changes = [(sql_gen.to_name_only_col(c, table), sql_gen.as_Value(v))
|
827
|
for c, v in changes]
|
828
|
|
829
|
if in_place:
|
830
|
assert cond == None
|
831
|
|
832
|
query = 'ALTER TABLE '+table.to_str(db)+'\n'
|
833
|
query += ',\n'.join(('ALTER COLUMN '+c.to_str(db)+' TYPE '
|
834
|
+db.col_info(sql_gen.with_default_table(c, table)).type
|
835
|
+'\nUSING '+v.to_str(db) for c, v in changes))
|
836
|
else:
|
837
|
query = 'UPDATE '+table.to_str(db)+'\nSET\n'
|
838
|
query += ',\n'.join((c.to_str(db)+' = '+v.to_str(db)
|
839
|
for c, v in changes))
|
840
|
if cond != None: query += '\nWHERE\n'+cond.to_str(db)
|
841
|
|
842
|
return query
|
843
|
|
844
|
def update(db, table, *args, **kw_args):
|
845
|
'''For params, see mk_update() and run_query()'''
|
846
|
recover = kw_args.pop('recover', None)
|
847
|
cacheable = kw_args.pop('cacheable', False)
|
848
|
log_level = kw_args.pop('log_level', 2)
|
849
|
|
850
|
cur = run_query(db, mk_update(db, table, *args, **kw_args), recover,
|
851
|
cacheable, log_level=log_level)
|
852
|
autoanalyze(db, table)
|
853
|
return cur
|
854
|
|
855
|
def last_insert_id(db):
|
856
|
module = util.root_module(db.db)
|
857
|
if module == 'psycopg2': return value(run_query(db, 'SELECT lastval()'))
|
858
|
elif module == 'MySQLdb': return db.insert_id()
|
859
|
else: return None
|
860
|
|
861
|
def mk_flatten_mapping(db, into, cols, preserve=[], as_items=False):
|
862
|
'''Creates a mapping from original column names (which may have collisions)
|
863
|
to names that will be distinct among the columns' tables.
|
864
|
This is meant to be used for several tables that are being joined together.
|
865
|
@param cols The columns to combine. Duplicates will be removed.
|
866
|
@param into The table for the new columns.
|
867
|
@param preserve [sql_gen.Col...] Columns not to rename. Note that these
|
868
|
columns will be included in the mapping even if they are not in cols.
|
869
|
The tables of the provided Col objects will be changed to into, so make
|
870
|
copies of them if you want to keep the original tables.
|
871
|
@param as_items Whether to return a list of dict items instead of a dict
|
872
|
@return dict(orig_col=new_col, ...)
|
873
|
* orig_col: sql_gen.Col(orig_col_name, orig_table)
|
874
|
* new_col: sql_gen.Col(orig_col_name, into)
|
875
|
* All mappings use the into table so its name can easily be
|
876
|
changed for all columns at once
|
877
|
'''
|
878
|
cols = lists.uniqify(cols)
|
879
|
|
880
|
items = []
|
881
|
for col in preserve:
|
882
|
orig_col = copy.copy(col)
|
883
|
col.table = into
|
884
|
items.append((orig_col, col))
|
885
|
preserve = set(preserve)
|
886
|
for col in cols:
|
887
|
if col not in preserve:
|
888
|
items.append((col, sql_gen.Col(str(col), into, col.srcs)))
|
889
|
|
890
|
if not as_items: items = dict(items)
|
891
|
return items
|
892
|
|
893
|
def flatten(db, into, joins, cols, limit=None, start=None, **kw_args):
|
894
|
'''For params, see mk_flatten_mapping()
|
895
|
@return See return value of mk_flatten_mapping()
|
896
|
'''
|
897
|
items = mk_flatten_mapping(db, into, cols, as_items=True, **kw_args)
|
898
|
cols = [sql_gen.NamedCol(new.name, old) for old, new in items]
|
899
|
run_query_into(db, mk_select(db, joins, cols, limit=limit, start=start),
|
900
|
into=into, add_indexes_=True)
|
901
|
return dict(items)
|
902
|
|
903
|
##### Database structure introspection
|
904
|
|
905
|
#### Tables
|
906
|
|
907
|
def tables(db, schema_like='public', table_like='%', exact=False):
|
908
|
if exact: compare = '='
|
909
|
else: compare = 'LIKE'
|
910
|
|
911
|
module = util.root_module(db.db)
|
912
|
if module == 'psycopg2':
|
913
|
conds = [('schemaname', sql_gen.CompareCond(schema_like, compare)),
|
914
|
('tablename', sql_gen.CompareCond(table_like, compare))]
|
915
|
return values(select(db, 'pg_tables', ['tablename'], conds,
|
916
|
order_by='tablename', log_level=4))
|
917
|
elif module == 'MySQLdb':
|
918
|
return values(run_query(db, 'SHOW TABLES LIKE '+db.esc_value(table_like)
|
919
|
, cacheable=True, log_level=4))
|
920
|
else: raise NotImplementedError("Can't list tables for "+module+' database')
|
921
|
|
922
|
def table_exists(db, table):
|
923
|
table = sql_gen.as_Table(table)
|
924
|
return list(tables(db, table.schema, table.name, exact=True)) != []
|
925
|
|
926
|
def table_row_count(db, table, recover=None):
|
927
|
return value(run_query(db, mk_select(db, table, [sql_gen.row_count],
|
928
|
order_by=None, start=0), recover=recover, log_level=3))
|
929
|
|
930
|
def table_cols(db, table, recover=None):
|
931
|
return list(col_names(select(db, table, limit=0, order_by=None,
|
932
|
recover=recover, log_level=4)))
|
933
|
|
934
|
def pkey(db, table, recover=None):
|
935
|
'''Assumed to be first column in table'''
|
936
|
return table_cols(db, table, recover)[0]
|
937
|
|
938
|
not_null_col = 'not_null_col'
|
939
|
|
940
|
def table_not_null_col(db, table, recover=None):
|
941
|
'''Name assumed to be the value of not_null_col. If not found, uses pkey.'''
|
942
|
if not_null_col in table_cols(db, table, recover): return not_null_col
|
943
|
else: return pkey(db, table, recover)
|
944
|
|
945
|
def index_cols(db, table, index):
|
946
|
'''Can also use this for UNIQUE constraints, because a UNIQUE index is
|
947
|
automatically created. When you don't know whether something is a UNIQUE
|
948
|
constraint or a UNIQUE index, use this function.'''
|
949
|
module = util.root_module(db.db)
|
950
|
if module == 'psycopg2':
|
951
|
return list(values(run_query(db, '''\
|
952
|
SELECT attname
|
953
|
FROM
|
954
|
(
|
955
|
SELECT attnum, attname
|
956
|
FROM pg_index
|
957
|
JOIN pg_class index ON index.oid = indexrelid
|
958
|
JOIN pg_class table_ ON table_.oid = indrelid
|
959
|
JOIN pg_attribute ON attrelid = indrelid AND attnum = ANY (indkey)
|
960
|
WHERE
|
961
|
table_.relname = '''+db.esc_value(table)+'''
|
962
|
AND index.relname = '''+db.esc_value(index)+'''
|
963
|
UNION
|
964
|
SELECT attnum, attname
|
965
|
FROM
|
966
|
(
|
967
|
SELECT
|
968
|
indrelid
|
969
|
, (regexp_matches(indexprs, E':varattno (\\\\d+)', 'g'))[1]::int
|
970
|
AS indkey
|
971
|
FROM pg_index
|
972
|
JOIN pg_class index ON index.oid = indexrelid
|
973
|
JOIN pg_class table_ ON table_.oid = indrelid
|
974
|
WHERE
|
975
|
table_.relname = '''+db.esc_value(table)+'''
|
976
|
AND index.relname = '''+db.esc_value(index)+'''
|
977
|
) s
|
978
|
JOIN pg_attribute ON attrelid = indrelid AND attnum = indkey
|
979
|
) s
|
980
|
ORDER BY attnum
|
981
|
'''
|
982
|
, cacheable=True, log_level=4)))
|
983
|
else: raise NotImplementedError("Can't list index columns for "+module+
|
984
|
' database')
|
985
|
|
986
|
def constraint_cols(db, table, constraint):
|
987
|
module = util.root_module(db.db)
|
988
|
if module == 'psycopg2':
|
989
|
return list(values(run_query(db, '''\
|
990
|
SELECT attname
|
991
|
FROM pg_constraint
|
992
|
JOIN pg_class ON pg_class.oid = conrelid
|
993
|
JOIN pg_attribute ON attrelid = conrelid AND attnum = ANY (conkey)
|
994
|
WHERE
|
995
|
relname = '''+db.esc_value(table)+'''
|
996
|
AND conname = '''+db.esc_value(constraint)+'''
|
997
|
ORDER BY attnum
|
998
|
'''
|
999
|
)))
|
1000
|
else: raise NotImplementedError("Can't list constraint columns for "+module+
|
1001
|
' database')
|
1002
|
|
1003
|
#### Functions
|
1004
|
|
1005
|
def function_exists(db, function):
|
1006
|
function = sql_gen.as_Function(function)
|
1007
|
|
1008
|
info_table = sql_gen.Table('routines', 'information_schema')
|
1009
|
conds = [('routine_name', function.name)]
|
1010
|
schema = function.schema
|
1011
|
if schema != None: conds.append(('routine_schema', schema))
|
1012
|
# Exclude trigger functions, since they cannot be called directly
|
1013
|
conds.append(('data_type', sql_gen.CompareCond('trigger', '!=')))
|
1014
|
|
1015
|
return list(values(select(db, info_table, ['routine_name'], conds,
|
1016
|
order_by='routine_schema', limit=1, log_level=4))) != []
|
1017
|
# TODO: order_by search_path schema order
|
1018
|
|
1019
|
##### Structural changes
|
1020
|
|
1021
|
#### Columns
|
1022
|
|
1023
|
def add_col(db, table, col, comment=None, **kw_args):
|
1024
|
'''
|
1025
|
@param col TypedCol Name may be versioned, so be sure to propagate any
|
1026
|
renaming back to any source column for the TypedCol.
|
1027
|
@param comment None|str SQL comment used to distinguish columns of the same
|
1028
|
name from each other when they contain different data, to allow the
|
1029
|
ADD COLUMN query to be cached. If not set, query will not be cached.
|
1030
|
'''
|
1031
|
assert isinstance(col, sql_gen.TypedCol)
|
1032
|
|
1033
|
while True:
|
1034
|
str_ = 'ALTER TABLE '+table.to_str(db)+' ADD COLUMN '+col.to_str(db)
|
1035
|
if comment != None: str_ += ' '+sql_gen.esc_comment(comment)
|
1036
|
|
1037
|
try:
|
1038
|
run_query(db, str_, recover=True, cacheable=True, **kw_args)
|
1039
|
break
|
1040
|
except DuplicateException:
|
1041
|
col.name = next_version(col.name)
|
1042
|
# try again with next version of name
|
1043
|
|
1044
|
def add_not_null(db, col):
|
1045
|
table = col.table
|
1046
|
col = sql_gen.to_name_only_col(col)
|
1047
|
run_query(db, 'ALTER TABLE '+table.to_str(db)+' ALTER COLUMN '
|
1048
|
+col.to_str(db)+' SET NOT NULL', cacheable=True, log_level=3)
|
1049
|
|
1050
|
row_num_col = '_row_num'
|
1051
|
|
1052
|
row_num_typed_col = sql_gen.TypedCol(row_num_col, 'serial', nullable=False,
|
1053
|
constraints='PRIMARY KEY')
|
1054
|
|
1055
|
def add_row_num(db, table):
|
1056
|
'''Adds a row number column to a table. Its name is in row_num_col. It will
|
1057
|
be the primary key.'''
|
1058
|
add_col(db, table, row_num_typed_col, log_level=3)
|
1059
|
|
1060
|
#### Indexes
|
1061
|
|
1062
|
def add_pkey(db, table, cols=None, recover=None):
|
1063
|
'''Adds a primary key.
|
1064
|
@param cols [sql_gen.Col,...] The columns in the primary key.
|
1065
|
Defaults to the first column in the table.
|
1066
|
@pre The table must not already have a primary key.
|
1067
|
'''
|
1068
|
table = sql_gen.as_Table(table)
|
1069
|
if cols == None: cols = [pkey(db, table, recover)]
|
1070
|
col_strs = [sql_gen.to_name_only_col(v).to_str(db) for v in cols]
|
1071
|
|
1072
|
run_query(db, 'ALTER TABLE '+table.to_str(db)+' ADD PRIMARY KEY ('
|
1073
|
+(', '.join(col_strs))+')', recover=True, cacheable=True, log_level=3,
|
1074
|
log_ignore_excs=(DuplicateException,))
|
1075
|
|
1076
|
def add_index(db, exprs, table=None, unique=False, ensure_not_null_=True):
|
1077
|
'''Adds an index on column(s) or expression(s) if it doesn't already exist.
|
1078
|
Currently, only function calls are supported as expressions.
|
1079
|
@param ensure_not_null_ If set, translates NULL values to sentinel values.
|
1080
|
This allows indexes to be used for comparisons where NULLs are equal.
|
1081
|
'''
|
1082
|
exprs = lists.mk_seq(exprs)
|
1083
|
|
1084
|
# Parse exprs
|
1085
|
old_exprs = exprs[:]
|
1086
|
exprs = []
|
1087
|
cols = []
|
1088
|
for i, expr in enumerate(old_exprs):
|
1089
|
expr = sql_gen.as_Col(expr, table)
|
1090
|
|
1091
|
# Handle nullable columns
|
1092
|
if ensure_not_null_:
|
1093
|
try: expr = ensure_not_null(db, expr)
|
1094
|
except KeyError: pass # unknown type, so just create plain index
|
1095
|
|
1096
|
# Extract col
|
1097
|
expr = copy.deepcopy(expr) # don't modify input!
|
1098
|
if isinstance(expr, sql_gen.FunctionCall):
|
1099
|
col = expr.args[0]
|
1100
|
expr = sql_gen.Expr(expr)
|
1101
|
else: col = expr
|
1102
|
assert isinstance(col, sql_gen.Col)
|
1103
|
|
1104
|
# Extract table
|
1105
|
if table == None:
|
1106
|
assert sql_gen.is_table_col(col)
|
1107
|
table = col.table
|
1108
|
|
1109
|
col.table = None
|
1110
|
|
1111
|
exprs.append(expr)
|
1112
|
cols.append(col)
|
1113
|
|
1114
|
table = sql_gen.as_Table(table)
|
1115
|
index = sql_gen.Table(str(sql_gen.Col(','.join(map(str, cols)), table)))
|
1116
|
|
1117
|
# Add index
|
1118
|
while True:
|
1119
|
str_ = 'CREATE'
|
1120
|
if unique: str_ += ' UNIQUE'
|
1121
|
str_ += ' INDEX '+index.to_str(db)+' ON '+table.to_str(db)+' ('+(
|
1122
|
', '.join((v.to_str(db) for v in exprs)))+')'
|
1123
|
|
1124
|
try:
|
1125
|
run_query(db, str_, recover=True, cacheable=True, log_level=3,
|
1126
|
log_ignore_excs=(DuplicateException,))
|
1127
|
break
|
1128
|
except DuplicateException:
|
1129
|
index.name = next_version(index.name)
|
1130
|
# try again with next version of name
|
1131
|
|
1132
|
def add_index_col(db, col, suffix, expr, nullable=True):
|
1133
|
if sql_gen.index_col(col) != None: return # already has index col
|
1134
|
|
1135
|
new_col = sql_gen.suffixed_col(col, suffix)
|
1136
|
|
1137
|
# Add column
|
1138
|
new_typed_col = sql_gen.TypedCol(new_col.name, db.col_info(col).type)
|
1139
|
add_col(db, col.table, new_typed_col, comment='src: '+repr(col),
|
1140
|
log_level=3)
|
1141
|
new_col.name = new_typed_col.name # propagate any renaming
|
1142
|
|
1143
|
update(db, col.table, [(new_col, expr)], in_place=True, cacheable=True,
|
1144
|
log_level=3)
|
1145
|
if not nullable: add_not_null(db, new_col)
|
1146
|
add_index(db, new_col)
|
1147
|
|
1148
|
col.table.index_cols[col.name] = new_col.name
|
1149
|
|
1150
|
# Controls when ensure_not_null() will use index columns
|
1151
|
not_null_index_cols_min_rows = 0 # rows; initially always use index columns
|
1152
|
|
1153
|
def ensure_not_null(db, col):
|
1154
|
'''For params, see sql_gen.ensure_not_null()'''
|
1155
|
expr = sql_gen.ensure_not_null(db, col)
|
1156
|
|
1157
|
# If a nullable column in a temp table, add separate index column instead.
|
1158
|
# Note that for small datasources, this adds 6-25% to the total import time.
|
1159
|
if (sql_gen.is_temp_col(col) and isinstance(expr, sql_gen.EnsureNotNull)
|
1160
|
and table_row_count(db, col.table) >= not_null_index_cols_min_rows):
|
1161
|
add_index_col(db, col, '::NOT NULL', expr, nullable=False)
|
1162
|
expr = sql_gen.index_col(col)
|
1163
|
|
1164
|
return expr
|
1165
|
|
1166
|
already_indexed = object() # tells add_indexes() the pkey has already been added
|
1167
|
|
1168
|
def add_indexes(db, table, has_pkey=True):
|
1169
|
'''Adds an index on all columns in a table.
|
1170
|
@param has_pkey bool|already_indexed Whether a pkey instead of a regular
|
1171
|
index should be added on the first column.
|
1172
|
* If already_indexed, the pkey is assumed to have already been added
|
1173
|
'''
|
1174
|
cols = table_cols(db, table)
|
1175
|
if has_pkey:
|
1176
|
if has_pkey is not already_indexed: add_pkey(db, table)
|
1177
|
cols = cols[1:]
|
1178
|
for col in cols: add_index(db, col, table)
|
1179
|
|
1180
|
#### Tables
|
1181
|
|
1182
|
### Maintenance
|
1183
|
|
1184
|
def analyze(db, table):
|
1185
|
table = sql_gen.as_Table(table)
|
1186
|
run_query(db, 'ANALYZE '+table.to_str(db), log_level=3)
|
1187
|
|
1188
|
def autoanalyze(db, table):
|
1189
|
if db.autoanalyze: analyze(db, table)
|
1190
|
|
1191
|
def vacuum(db, table):
|
1192
|
table = sql_gen.as_Table(table)
|
1193
|
db.with_autocommit(lambda: run_query(db, 'VACUUM ANALYZE '+table.to_str(db),
|
1194
|
log_level=3))
|
1195
|
|
1196
|
### Lifecycle
|
1197
|
|
1198
|
def drop_table(db, table):
|
1199
|
table = sql_gen.as_Table(table)
|
1200
|
return run_query(db, 'DROP TABLE IF EXISTS '+table.to_str(db)+' CASCADE')
|
1201
|
|
1202
|
def create_table(db, table, cols=[], has_pkey=True, col_indexes=True,
|
1203
|
like=None):
|
1204
|
'''Creates a table.
|
1205
|
@param cols [sql_gen.TypedCol,...] The column names and types
|
1206
|
@param has_pkey If set, the first column becomes the primary key.
|
1207
|
@param col_indexes bool|[ref]
|
1208
|
* If True, indexes will be added on all non-pkey columns.
|
1209
|
* If a list reference, [0] will be set to a function to do this.
|
1210
|
This can be used to delay index creation until the table is populated.
|
1211
|
'''
|
1212
|
table = sql_gen.as_Table(table)
|
1213
|
|
1214
|
if like != None:
|
1215
|
cols = [sql_gen.CustomCode('LIKE '+like.to_str(db)+' INCLUDING ALL')
|
1216
|
]+cols
|
1217
|
if has_pkey:
|
1218
|
cols[0] = pkey = copy.copy(cols[0]) # don't modify input!
|
1219
|
pkey.constraints = 'PRIMARY KEY'
|
1220
|
|
1221
|
temp = table.is_temp and not db.debug_temp
|
1222
|
# temp tables permanent in debug_temp mode
|
1223
|
|
1224
|
# Create table
|
1225
|
while True:
|
1226
|
str_ = 'CREATE'
|
1227
|
if temp: str_ += ' TEMP'
|
1228
|
str_ += ' TABLE '+table.to_str(db)+' (\n'
|
1229
|
str_ += '\n, '.join(c.to_str(db) for c in cols)
|
1230
|
str_ += '\n);'
|
1231
|
|
1232
|
try:
|
1233
|
run_query(db, str_, recover=True, cacheable=True, log_level=2,
|
1234
|
log_ignore_excs=(DuplicateException,))
|
1235
|
break
|
1236
|
except DuplicateException:
|
1237
|
table.name = next_version(table.name)
|
1238
|
# try again with next version of name
|
1239
|
|
1240
|
# Add indexes
|
1241
|
if has_pkey: has_pkey = already_indexed
|
1242
|
def add_indexes_(): add_indexes(db, table, has_pkey)
|
1243
|
if isinstance(col_indexes, list): col_indexes[0] = add_indexes_ # defer
|
1244
|
elif col_indexes: add_indexes_() # add now
|
1245
|
|
1246
|
def copy_table_struct(db, src, dest):
|
1247
|
'''Creates a structure-only copy of a table. (Does not copy data.)'''
|
1248
|
create_table(db, dest, has_pkey=False, col_indexes=False, like=src)
|
1249
|
|
1250
|
### Data
|
1251
|
|
1252
|
def truncate(db, table, schema='public', **kw_args):
|
1253
|
'''For params, see run_query()'''
|
1254
|
table = sql_gen.as_Table(table, schema)
|
1255
|
return run_query(db, 'TRUNCATE '+table.to_str(db)+' CASCADE', **kw_args)
|
1256
|
|
1257
|
def empty_temp(db, tables):
|
1258
|
if db.debug_temp: return # leave temp tables there for debugging
|
1259
|
tables = lists.mk_seq(tables)
|
1260
|
for table in tables: truncate(db, table, log_level=3)
|
1261
|
|
1262
|
def empty_db(db, schema='public', **kw_args):
|
1263
|
'''For kw_args, see tables()'''
|
1264
|
for table in tables(db, schema, **kw_args): truncate(db, table, schema)
|
1265
|
|
1266
|
def distinct_table(db, table, distinct_on):
|
1267
|
'''Creates a copy of a temp table which is distinct on the given columns.
|
1268
|
The old and new tables will both get an index on these columns, to
|
1269
|
facilitate merge joins.
|
1270
|
@param distinct_on If empty, creates a table with one row. This is useful if
|
1271
|
your distinct_on columns are all literal values.
|
1272
|
@return The new table.
|
1273
|
'''
|
1274
|
new_table = sql_gen.suffixed_table(table, '_distinct')
|
1275
|
|
1276
|
copy_table_struct(db, table, new_table)
|
1277
|
|
1278
|
limit = None
|
1279
|
if distinct_on == []: limit = 1 # one sample row
|
1280
|
else:
|
1281
|
add_index(db, distinct_on, new_table, unique=True)
|
1282
|
add_index(db, distinct_on, table) # for join optimization
|
1283
|
|
1284
|
insert_select(db, new_table, None, mk_select(db, table, start=0,
|
1285
|
limit=limit), ignore=True)
|
1286
|
analyze(db, new_table)
|
1287
|
|
1288
|
return new_table
|