Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.11/site-packages/sqlalchemy/sql/_elements_constructors.py: 62%

Shortcuts on this page

r m x   toggle line displays

j k   next/prev highlighted chunk

0   (zero) top of page

1   (one) first highlighted chunk

171 statements  

1# sql/_elements_constructors.py 

2# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors 

3# <see AUTHORS file> 

4# 

5# This module is part of SQLAlchemy and is released under 

6# the MIT License: https://www.opensource.org/licenses/mit-license.php 

7 

8from __future__ import annotations 

9 

10import typing 

11from typing import Any 

12from typing import Callable 

13from typing import Literal 

14from typing import Mapping 

15from typing import Optional 

16from typing import overload 

17from typing import Sequence 

18from typing import Tuple as typing_Tuple 

19from typing import TYPE_CHECKING 

20from typing import TypeVar 

21from typing import Union 

22 

23from . import coercions 

24from . import operators 

25from . import roles 

26from .base import _NoArg 

27from .coercions import _document_text_coercion 

28from .elements import AggregateOrderBy 

29from .elements import BindParameter 

30from .elements import BooleanClauseList 

31from .elements import Case 

32from .elements import Cast 

33from .elements import CollationClause 

34from .elements import CollectionAggregate 

35from .elements import ColumnClause 

36from .elements import ColumnElement 

37from .elements import DMLTargetCopy 

38from .elements import Extract 

39from .elements import False_ 

40from .elements import FunctionFilter 

41from .elements import Label 

42from .elements import Null 

43from .elements import OrderByList 

44from .elements import Over 

45from .elements import TextClause 

46from .elements import True_ 

47from .elements import TryCast 

48from .elements import TString 

49from .elements import Tuple 

50from .elements import TypeCoerce 

51from .elements import UnaryExpression 

52from .elements import WithinGroup 

53from .functions import FunctionElement 

54 

55if typing.TYPE_CHECKING: 

56 from string.templatelib import Template 

57 

58 from ._typing import _ByArgument 

59 from ._typing import _ColumnExpressionArgument 

60 from ._typing import _ColumnExpressionOrLiteralArgument 

61 from ._typing import _ColumnExpressionOrStrLabelArgument 

62 from ._typing import _OnlyColumnArgument 

63 from ._typing import _TypeEngineArgument 

64 from .elements import _FrameIntTuple 

65 from .elements import BinaryExpression 

66 from .elements import FrameClause 

67 from .selectable import FromClause 

68 from .type_api import TypeEngine 

69 

70_T = TypeVar("_T") 

71 

72 

73def all_(expr: _ColumnExpressionArgument[_T]) -> CollectionAggregate[bool]: 

74 """Produce an ALL expression. 

75 

76 For dialects such as that of PostgreSQL, this operator applies 

77 to usage of the :class:`_types.ARRAY` datatype, for that of 

78 MySQL, it may apply to a subquery. e.g.:: 

79 

80 # renders on PostgreSQL: 

81 # '5 = ALL (somearray)' 

82 expr = 5 == all_(mytable.c.somearray) 

83 

84 # renders on MySQL: 

85 # '5 = ALL (SELECT value FROM table)' 

86 expr = 5 == all_(select(table.c.value)) 

87 

88 Comparison to NULL may work using ``None``:: 

89 

90 None == all_(mytable.c.somearray) 

91 

92 The any_() / all_() operators also feature a special "operand flipping" 

93 behavior such that if any_() / all_() are used on the left side of a 

94 comparison using a standalone operator such as ``==``, ``!=``, etc. 

95 (not including operator methods such as 

96 :meth:`_sql.ColumnOperators.is_`) the rendered expression is flipped:: 

97 

98 # would render '5 = ALL (column)` 

99 all_(mytable.c.column) == 5 

100 

101 Or with ``None``, which note will not perform 

102 the usual step of rendering "IS" as is normally the case for NULL:: 

103 

104 # would render 'NULL = ALL(somearray)' 

105 all_(mytable.c.somearray) == None 

106 

107 The column-level :meth:`_sql.ColumnElement.all_` method (not to be 

108 confused with the deprecated :class:`_types.ARRAY` level 

109 :meth:`_types.ARRAY.Comparator.all`) is shorthand for 

110 ``all_(col)``:: 

111 

112 5 == mytable.c.somearray.all_() 

113 

114 .. seealso:: 

115 

116 :meth:`_sql.ColumnOperators.all_` 

117 

118 :func:`_expression.any_` 

119 

120 """ 

121 if isinstance(expr, operators.ColumnOperators): 

122 return expr.all_() 

123 else: 

124 return CollectionAggregate._create_all(expr) 

125 

126 

127def and_( # type: ignore[empty-body] 

128 initial_clause: Union[Literal[True], _ColumnExpressionArgument[bool]], 

129 *clauses: _ColumnExpressionArgument[bool], 

130) -> ColumnElement[bool]: 

131 r"""Produce a conjunction of expressions joined by ``AND``. 

132 

133 E.g.:: 

134 

135 from sqlalchemy import and_ 

136 

137 stmt = select(users_table).where( 

138 and_(users_table.c.name == "wendy", users_table.c.enrolled == True) 

139 ) 

140 

141 The :func:`.and_` conjunction is also available using the 

142 Python ``&`` operator (though note that compound expressions 

143 need to be parenthesized in order to function with Python 

144 operator precedence behavior):: 

145 

146 stmt = select(users_table).where( 

147 (users_table.c.name == "wendy") & (users_table.c.enrolled == True) 

148 ) 

149 

150 The :func:`.and_` operation is also implicit in some cases; 

151 the :meth:`_expression.Select.where` 

152 method for example can be invoked multiple 

153 times against a statement, which will have the effect of each 

154 clause being combined using :func:`.and_`:: 

155 

156 stmt = ( 

157 select(users_table) 

158 .where(users_table.c.name == "wendy") 

159 .where(users_table.c.enrolled == True) 

160 ) 

161 

162 The :func:`.and_` construct must be given at least one positional 

163 argument in order to be valid; a :func:`.and_` construct with no 

164 arguments is ambiguous. To produce an "empty" or dynamically 

165 generated :func:`.and_` expression, from a given list of expressions, 

166 a "default" element of :func:`_sql.true` (or just ``True``) should be 

167 specified:: 

168 

169 from sqlalchemy import true 

170 

171 criteria = and_(true(), *expressions) 

172 

173 The above expression will compile to SQL as the expression ``true`` 

174 or ``1 = 1``, depending on backend, if no other expressions are 

175 present. If expressions are present, then the :func:`_sql.true` value is 

176 ignored as it does not affect the outcome of an AND expression that 

177 has other elements. 

178 

179 .. deprecated:: 1.4 The :func:`.and_` element now requires that at 

180 least one argument is passed; creating the :func:`.and_` construct 

181 with no arguments is deprecated, and will emit a deprecation warning 

182 while continuing to produce a blank SQL string. 

183 

184 .. seealso:: 

185 

186 :func:`.or_` 

187 

188 """ 

189 ... 

190 

191 

192if not TYPE_CHECKING: 

193 # handle deprecated case which allows zero-arguments 

194 def and_(*clauses): # noqa: F811 

195 r"""Produce a conjunction of expressions joined by ``AND``. 

196 

197 E.g.:: 

198 

199 from sqlalchemy import and_ 

200 

201 stmt = select(users_table).where( 

202 and_(users_table.c.name == "wendy", users_table.c.enrolled == True) 

203 ) 

204 

205 The :func:`.and_` conjunction is also available using the 

206 Python ``&`` operator (though note that compound expressions 

207 need to be parenthesized in order to function with Python 

208 operator precedence behavior):: 

209 

210 stmt = select(users_table).where( 

211 (users_table.c.name == "wendy") & (users_table.c.enrolled == True) 

212 ) 

213 

214 The :func:`.and_` operation is also implicit in some cases; 

215 the :meth:`_expression.Select.where` 

216 method for example can be invoked multiple 

217 times against a statement, which will have the effect of each 

218 clause being combined using :func:`.and_`:: 

219 

220 stmt = ( 

221 select(users_table) 

222 .where(users_table.c.name == "wendy") 

223 .where(users_table.c.enrolled == True) 

224 ) 

225 

226 The :func:`.and_` construct must be given at least one positional 

227 argument in order to be valid; a :func:`.and_` construct with no 

228 arguments is ambiguous. To produce an "empty" or dynamically 

229 generated :func:`.and_` expression, from a given list of expressions, 

230 a "default" element of :func:`_sql.true` (or just ``True``) should be 

231 specified:: 

232 

233 from sqlalchemy import true 

234 

235 criteria = and_(true(), *expressions) 

236 

237 The above expression will compile to SQL as the expression ``true`` 

238 or ``1 = 1``, depending on backend, if no other expressions are 

239 present. If expressions are present, then the :func:`_sql.true` value 

240 is ignored as it does not affect the outcome of an AND expression that 

241 has other elements. 

242 

243 .. deprecated:: 1.4 The :func:`.and_` element now requires that at 

244 least one argument is passed; creating the :func:`.and_` construct 

245 with no arguments is deprecated, and will emit a deprecation warning 

246 while continuing to produce a blank SQL string. 

247 

248 .. seealso:: 

249 

250 :func:`.or_` 

251 

252 """ # noqa: E501 

253 return BooleanClauseList.and_(*clauses) 

254 

255 

256def any_(expr: _ColumnExpressionArgument[_T]) -> CollectionAggregate[bool]: 

257 """Produce an ANY expression. 

258 

259 For dialects such as that of PostgreSQL, this operator applies 

260 to usage of the :class:`_types.ARRAY` datatype, for that of 

261 MySQL, it may apply to a subquery. e.g.:: 

262 

263 # renders on PostgreSQL: 

264 # '5 = ANY (somearray)' 

265 expr = 5 == any_(mytable.c.somearray) 

266 

267 # renders on MySQL: 

268 # '5 = ANY (SELECT value FROM table)' 

269 expr = 5 == any_(select(table.c.value)) 

270 

271 Comparison to NULL may work using ``None`` or :func:`_sql.null`:: 

272 

273 None == any_(mytable.c.somearray) 

274 

275 The any_() / all_() operators also feature a special "operand flipping" 

276 behavior such that if any_() / all_() are used on the left side of a 

277 comparison using a standalone operator such as ``==``, ``!=``, etc. 

278 (not including operator methods such as 

279 :meth:`_sql.ColumnOperators.is_`) the rendered expression is flipped:: 

280 

281 # would render '5 = ANY (column)` 

282 any_(mytable.c.column) == 5 

283 

284 Or with ``None``, which note will not perform 

285 the usual step of rendering "IS" as is normally the case for NULL:: 

286 

287 # would render 'NULL = ANY(somearray)' 

288 any_(mytable.c.somearray) == None 

289 

290 The column-level :meth:`_sql.ColumnElement.any_` method (not to be 

291 confused with the deprecated :class:`_types.ARRAY` level 

292 :meth:`_types.ARRAY.Comparator.any`) is shorthand for 

293 ``any_(col)``:: 

294 

295 5 = mytable.c.somearray.any_() 

296 

297 .. seealso:: 

298 

299 :meth:`_sql.ColumnOperators.any_` 

300 

301 :func:`_expression.all_` 

302 

303 """ 

304 if isinstance(expr, operators.ColumnOperators): 

305 return expr.any_() 

306 else: 

307 return CollectionAggregate._create_any(expr) 

308 

309 

310@overload 

311def asc( 

312 column: Union[str, "ColumnElement[_T]"], 

313) -> UnaryExpression[_T]: ... 

314 

315 

316@overload 

317def asc( 

318 column: _ColumnExpressionOrStrLabelArgument[_T], 

319) -> Union[OrderByList, UnaryExpression[_T]]: ... 

320 

321 

322def asc( 

323 column: _ColumnExpressionOrStrLabelArgument[_T], 

324) -> Union[OrderByList, UnaryExpression[_T]]: 

325 """Produce an ascending ``ORDER BY`` clause element. 

326 

327 e.g.:: 

328 

329 from sqlalchemy import asc 

330 

331 stmt = select(users_table).order_by(asc(users_table.c.name)) 

332 

333 will produce SQL as: 

334 

335 .. sourcecode:: sql 

336 

337 SELECT id, name FROM user ORDER BY name ASC 

338 

339 The :func:`.asc` function is a standalone version of the 

340 :meth:`_expression.ColumnElement.asc` 

341 method available on all SQL expressions, 

342 e.g.:: 

343 

344 

345 stmt = select(users_table).order_by(users_table.c.name.asc()) 

346 

347 :param column: A :class:`_expression.ColumnElement` (e.g. 

348 scalar SQL expression) 

349 with which to apply the :func:`.asc` operation. 

350 

351 .. seealso:: 

352 

353 :func:`.desc` 

354 

355 :func:`.nulls_first` 

356 

357 :func:`.nulls_last` 

358 

359 :meth:`_expression.Select.order_by` 

360 

361 """ 

362 

363 if isinstance(column, operators.OrderingOperators): 

364 return column.asc() # type: ignore[unused-ignore] 

365 else: 

366 return UnaryExpression._create_asc(column) 

367 

368 

369def collate( 

370 expression: _ColumnExpressionArgument[str], collation: str 

371) -> BinaryExpression[str]: 

372 """Return the clause ``expression COLLATE collation``. 

373 

374 e.g.:: 

375 

376 collate(mycolumn, "utf8_bin") 

377 

378 produces: 

379 

380 .. sourcecode:: sql 

381 

382 mycolumn COLLATE utf8_bin 

383 

384 The collation expression is also quoted if it is a case sensitive 

385 identifier, e.g. contains uppercase characters. 

386 

387 """ 

388 if isinstance(expression, operators.ColumnOperators): 

389 return expression.collate(collation) # type: ignore 

390 else: 

391 return CollationClause._create_collation_expression( 

392 expression, collation 

393 ) 

394 

395 

396def between( 

397 expr: _ColumnExpressionOrLiteralArgument[_T], 

398 lower_bound: Any, 

399 upper_bound: Any, 

400 symmetric: bool = False, 

401) -> BinaryExpression[bool]: 

402 """Produce a ``BETWEEN`` predicate clause. 

403 

404 E.g.:: 

405 

406 from sqlalchemy import between 

407 

408 stmt = select(users_table).where(between(users_table.c.id, 5, 7)) 

409 

410 Would produce SQL resembling: 

411 

412 .. sourcecode:: sql 

413 

414 SELECT id, name FROM user WHERE id BETWEEN :id_1 AND :id_2 

415 

416 The :func:`.between` function is a standalone version of the 

417 :meth:`_expression.ColumnElement.between` method available on all 

418 SQL expressions, as in:: 

419 

420 stmt = select(users_table).where(users_table.c.id.between(5, 7)) 

421 

422 All arguments passed to :func:`.between`, including the left side 

423 column expression, are coerced from Python scalar values if a 

424 the value is not a :class:`_expression.ColumnElement` subclass. 

425 For example, 

426 three fixed values can be compared as in:: 

427 

428 print(between(5, 3, 7)) 

429 

430 Which would produce:: 

431 

432 :param_1 BETWEEN :param_2 AND :param_3 

433 

434 :param expr: a column expression, typically a 

435 :class:`_expression.ColumnElement` 

436 instance or alternatively a Python scalar expression to be coerced 

437 into a column expression, serving as the left side of the ``BETWEEN`` 

438 expression. 

439 

440 :param lower_bound: a column or Python scalar expression serving as the 

441 lower bound of the right side of the ``BETWEEN`` expression. 

442 

443 :param upper_bound: a column or Python scalar expression serving as the 

444 upper bound of the right side of the ``BETWEEN`` expression. 

445 

446 :param symmetric: if True, will render " BETWEEN SYMMETRIC ". Note 

447 that not all databases support this syntax. 

448 

449 .. seealso:: 

450 

451 :meth:`_expression.ColumnElement.between` 

452 

453 """ 

454 col_expr = coercions.expect(roles.ExpressionElementRole, expr) 

455 return col_expr.between(lower_bound, upper_bound, symmetric=symmetric) 

456 

457 

458def outparam( 

459 key: str, type_: Optional[TypeEngine[_T]] = None 

460) -> BindParameter[_T]: 

461 """Create an 'OUT' parameter for usage in functions (stored procedures), 

462 for databases which support them. 

463 

464 The ``outparam`` can be used like a regular function parameter. 

465 The "output" value will be available from the 

466 :class:`~sqlalchemy.engine.CursorResult` object via its ``out_parameters`` 

467 attribute, which returns a dictionary containing the values. 

468 

469 """ 

470 return BindParameter(key, None, type_=type_, unique=False, isoutparam=True) 

471 

472 

473@overload 

474def not_(clause: BinaryExpression[_T]) -> BinaryExpression[_T]: ... 

475 

476 

477@overload 

478def not_(clause: _ColumnExpressionArgument[_T]) -> ColumnElement[_T]: ... 

479 

480 

481def not_(clause: _ColumnExpressionArgument[_T]) -> ColumnElement[_T]: 

482 """Return a negation of the given clause, i.e. ``NOT(clause)``. 

483 

484 The ``~`` operator is also overloaded on all 

485 :class:`_expression.ColumnElement` subclasses to produce the 

486 same result. 

487 

488 """ 

489 

490 return coercions.expect(roles.ExpressionElementRole, clause).__invert__() 

491 

492 

493def from_dml_column(column: _OnlyColumnArgument[_T]) -> DMLTargetCopy[_T]: 

494 r"""A placeholder that may be used in compiled INSERT or UPDATE expressions 

495 to refer to the SQL expression or value being applied to another column. 

496 

497 Given a table such as:: 

498 

499 t = Table( 

500 "t", 

501 MetaData(), 

502 Column("x", Integer), 

503 Column("y", Integer), 

504 ) 

505 

506 The :func:`_sql.from_dml_column` construct allows automatic copying 

507 of an expression assigned to a different column to be re-used:: 

508 

509 >>> stmt = t.insert().values(x=func.foobar(3), y=from_dml_column(t.c.x) + 5) 

510 >>> print(stmt) 

511 INSERT INTO t (x, y) VALUES (foobar(:foobar_1), (foobar(:foobar_1) + :param_1)) 

512 

513 The :func:`_sql.from_dml_column` construct is intended to be useful primarily 

514 with event-based hooks such as those used by ORM hybrids. 

515 

516 .. seealso:: 

517 

518 :ref:`hybrid_bulk_update` 

519 

520 .. versionadded:: 2.1 

521 

522 

523 """ # noqa: E501 

524 

525 return DMLTargetCopy(column) 

526 

527 

528def bindparam( 

529 key: Optional[str], 

530 value: Any = _NoArg.NO_ARG, 

531 type_: Optional[_TypeEngineArgument[_T]] = None, 

532 unique: bool = False, 

533 required: Union[bool, Literal[_NoArg.NO_ARG]] = _NoArg.NO_ARG, 

534 quote: Optional[bool] = None, 

535 callable_: Optional[Callable[[], Any]] = None, 

536 expanding: bool = False, 

537 isoutparam: bool = False, 

538 literal_execute: bool = False, 

539) -> BindParameter[_T]: 

540 r"""Produce a "bound expression". 

541 

542 The return value is an instance of :class:`.BindParameter`; this 

543 is a :class:`_expression.ColumnElement` 

544 subclass which represents a so-called 

545 "placeholder" value in a SQL expression, the value of which is 

546 supplied at the point at which the statement in executed against a 

547 database connection. 

548 

549 In SQLAlchemy, the :func:`.bindparam` construct has 

550 the ability to carry along the actual value that will be ultimately 

551 used at expression time. In this way, it serves not just as 

552 a "placeholder" for eventual population, but also as a means of 

553 representing so-called "unsafe" values which should not be rendered 

554 directly in a SQL statement, but rather should be passed along 

555 to the :term:`DBAPI` as values which need to be correctly escaped 

556 and potentially handled for type-safety. 

557 

558 When using :func:`.bindparam` explicitly, the use case is typically 

559 one of traditional deferment of parameters; the :func:`.bindparam` 

560 construct accepts a name which can then be referred to at execution 

561 time:: 

562 

563 from sqlalchemy import bindparam 

564 

565 stmt = select(users_table).where( 

566 users_table.c.name == bindparam("username") 

567 ) 

568 

569 The above statement, when rendered, will produce SQL similar to: 

570 

571 .. sourcecode:: sql 

572 

573 SELECT id, name FROM user WHERE name = :username 

574 

575 In order to populate the value of ``:username`` above, the value 

576 would typically be applied at execution time to a method 

577 like :meth:`_engine.Connection.execute`:: 

578 

579 result = connection.execute(stmt, {"username": "wendy"}) 

580 

581 Explicit use of :func:`.bindparam` is also common when producing 

582 UPDATE or DELETE statements that are to be invoked multiple times, 

583 where the WHERE criterion of the statement is to change on each 

584 invocation, such as:: 

585 

586 stmt = ( 

587 users_table.update() 

588 .where(user_table.c.name == bindparam("username")) 

589 .values(fullname=bindparam("fullname")) 

590 ) 

591 

592 connection.execute( 

593 stmt, 

594 [ 

595 {"username": "wendy", "fullname": "Wendy Smith"}, 

596 {"username": "jack", "fullname": "Jack Jones"}, 

597 ], 

598 ) 

599 

600 SQLAlchemy's Core expression system makes wide use of 

601 :func:`.bindparam` in an implicit sense. It is typical that Python 

602 literal values passed to virtually all SQL expression functions are 

603 coerced into fixed :func:`.bindparam` constructs. For example, given 

604 a comparison operation such as:: 

605 

606 expr = users_table.c.name == "Wendy" 

607 

608 The above expression will produce a :class:`.BinaryExpression` 

609 construct, where the left side is the :class:`_schema.Column` object 

610 representing the ``name`` column, and the right side is a 

611 :class:`.BindParameter` representing the literal value:: 

612 

613 print(repr(expr.right)) 

614 BindParameter("%(4327771088 name)s", "Wendy", type_=String()) 

615 

616 The expression above will render SQL such as: 

617 

618 .. sourcecode:: sql 

619 

620 user.name = :name_1 

621 

622 Where the ``:name_1`` parameter name is an anonymous name. The 

623 actual string ``Wendy`` is not in the rendered string, but is carried 

624 along where it is later used within statement execution. If we 

625 invoke a statement like the following:: 

626 

627 stmt = select(users_table).where(users_table.c.name == "Wendy") 

628 result = connection.execute(stmt) 

629 

630 We would see SQL logging output as: 

631 

632 .. sourcecode:: sql 

633 

634 SELECT "user".id, "user".name 

635 FROM "user" 

636 WHERE "user".name = %(name_1)s 

637 {'name_1': 'Wendy'} 

638 

639 Above, we see that ``Wendy`` is passed as a parameter to the database, 

640 while the placeholder ``:name_1`` is rendered in the appropriate form 

641 for the target database, in this case the PostgreSQL database. 

642 

643 Similarly, :func:`.bindparam` is invoked automatically when working 

644 with :term:`CRUD` statements as far as the "VALUES" portion is 

645 concerned. The :func:`_expression.insert` construct produces an 

646 ``INSERT`` expression which will, at statement execution time, generate 

647 bound placeholders based on the arguments passed, as in:: 

648 

649 stmt = users_table.insert() 

650 result = connection.execute(stmt, {"name": "Wendy"}) 

651 

652 The above will produce SQL output as: 

653 

654 .. sourcecode:: sql 

655 

656 INSERT INTO "user" (name) VALUES (%(name)s) 

657 {'name': 'Wendy'} 

658 

659 The :class:`_expression.Insert` construct, at 

660 compilation/execution time, rendered a single :func:`.bindparam` 

661 mirroring the column name ``name`` as a result of the single ``name`` 

662 parameter we passed to the :meth:`_engine.Connection.execute` method. 

663 

664 :param key: 

665 the key (e.g. the name) for this bind param. 

666 Will be used in the generated 

667 SQL statement for dialects that use named parameters. This 

668 value may be modified when part of a compilation operation, 

669 if other :class:`BindParameter` objects exist with the same 

670 key, or if its length is too long and truncation is 

671 required. 

672 

673 If omitted, an "anonymous" name is generated for the bound parameter; 

674 when given a value to bind, the end result is equivalent to calling upon 

675 the :func:`.literal` function with a value to bind, particularly 

676 if the :paramref:`.bindparam.unique` parameter is also provided. 

677 

678 :param value: 

679 Initial value for this bind param. Will be used at statement 

680 execution time as the value for this parameter passed to the 

681 DBAPI, if no other value is indicated to the statement execution 

682 method for this particular parameter name. Defaults to ``None``. 

683 

684 :param callable\_: 

685 A callable function that takes the place of "value". The function 

686 will be called at statement execution time to determine the 

687 ultimate value. Used for scenarios where the actual bind 

688 value cannot be determined at the point at which the clause 

689 construct is created, but embedded bind values are still desirable. 

690 

691 :param type\_: 

692 A :class:`.TypeEngine` class or instance representing an optional 

693 datatype for this :func:`.bindparam`. If not passed, a type 

694 may be determined automatically for the bind, based on the given 

695 value; for example, trivial Python types such as ``str``, 

696 ``int``, ``bool`` 

697 may result in the :class:`.String`, :class:`.Integer` or 

698 :class:`.Boolean` types being automatically selected. 

699 

700 The type of a :func:`.bindparam` is significant especially in that 

701 the type will apply pre-processing to the value before it is 

702 passed to the database. For example, a :func:`.bindparam` which 

703 refers to a datetime value, and is specified as holding the 

704 :class:`.DateTime` type, may apply conversion needed to the 

705 value (such as stringification on SQLite) before passing the value 

706 to the database. 

707 

708 :param unique: 

709 if True, the key name of this :class:`.BindParameter` will be 

710 modified if another :class:`.BindParameter` of the same name 

711 already has been located within the containing 

712 expression. This flag is used generally by the internals 

713 when producing so-called "anonymous" bound expressions, it 

714 isn't generally applicable to explicitly-named :func:`.bindparam` 

715 constructs. 

716 

717 :param required: 

718 If ``True``, a value is required at execution time. If not passed, 

719 it defaults to ``True`` if neither :paramref:`.bindparam.value` 

720 or :paramref:`.bindparam.callable` were passed. If either of these 

721 parameters are present, then :paramref:`.bindparam.required` 

722 defaults to ``False``. 

723 

724 :param quote: 

725 True if this parameter name requires quoting and is not 

726 currently known as a SQLAlchemy reserved word; this currently 

727 only applies to the Oracle Database backends, where bound names must 

728 sometimes be quoted. 

729 

730 :param isoutparam: 

731 if True, the parameter should be treated like a stored procedure 

732 "OUT" parameter. This applies to backends such as Oracle Database which 

733 support OUT parameters. 

734 

735 :param expanding: 

736 if True, this parameter will be treated as an "expanding" parameter 

737 at execution time; the parameter value is expected to be a sequence, 

738 rather than a scalar value, and the string SQL statement will 

739 be transformed on a per-execution basis to accommodate the sequence 

740 with a variable number of parameter slots passed to the DBAPI. 

741 This is to allow statement caching to be used in conjunction with 

742 an IN clause. 

743 

744 .. seealso:: 

745 

746 :meth:`.ColumnOperators.in_` 

747 

748 :ref:`baked_in` - with baked queries 

749 

750 .. note:: The "expanding" feature does not support "executemany"- 

751 style parameter sets. 

752 

753 :param literal_execute: 

754 if True, the bound parameter will be rendered in the compile phase 

755 with a special "POSTCOMPILE" token, and the SQLAlchemy compiler will 

756 render the final value of the parameter into the SQL statement at 

757 statement execution time, omitting the value from the parameter 

758 dictionary / list passed to DBAPI ``cursor.execute()``. This 

759 produces a similar effect as that of using the ``literal_binds``, 

760 compilation flag, however takes place as the statement is sent to 

761 the DBAPI ``cursor.execute()`` method, rather than when the statement 

762 is compiled. The primary use of this 

763 capability is for rendering LIMIT / OFFSET clauses for database 

764 drivers that can't accommodate for bound parameters in these 

765 contexts, while allowing SQL constructs to be cacheable at the 

766 compilation level. 

767 

768 .. versionadded:: 1.4 Added "post compile" bound parameters 

769 

770 .. seealso:: 

771 

772 :ref:`change_4808`. 

773 

774 .. seealso:: 

775 

776 :ref:`tutorial_sending_parameters` - in the 

777 :ref:`unified_tutorial` 

778 

779 

780 """ 

781 return BindParameter( 

782 key, 

783 value, 

784 type_, 

785 unique, 

786 required, 

787 quote, 

788 callable_, 

789 expanding, 

790 isoutparam, 

791 literal_execute, 

792 ) 

793 

794 

795def case( 

796 *whens: Union[ 

797 typing_Tuple[_ColumnExpressionArgument[bool], Any], Mapping[Any, Any] 

798 ], 

799 value: Optional[Any] = None, 

800 else_: Optional[Any] = None, 

801) -> Case[Any]: 

802 r"""Produce a ``CASE`` expression. 

803 

804 The ``CASE`` construct in SQL is a conditional object that 

805 acts somewhat analogously to an "if/then" construct in other 

806 languages. It returns an instance of :class:`.Case`. 

807 

808 :func:`.case` in its usual form is passed a series of "when" 

809 constructs, that is, a list of conditions and results as tuples:: 

810 

811 from sqlalchemy import case 

812 

813 stmt = select(users_table).where( 

814 case( 

815 (users_table.c.name == "wendy", "W"), 

816 (users_table.c.name == "jack", "J"), 

817 else_="E", 

818 ) 

819 ) 

820 

821 The above statement will produce SQL resembling: 

822 

823 .. sourcecode:: sql 

824 

825 SELECT id, name FROM user 

826 WHERE CASE 

827 WHEN (name = :name_1) THEN :param_1 

828 WHEN (name = :name_2) THEN :param_2 

829 ELSE :param_3 

830 END 

831 

832 When simple equality expressions of several values against a single 

833 parent column are needed, :func:`.case` also has a "shorthand" format 

834 used via the 

835 :paramref:`.case.value` parameter, which is passed a column 

836 expression to be compared. In this form, the :paramref:`.case.whens` 

837 parameter is passed as a dictionary containing expressions to be 

838 compared against keyed to result expressions. The statement below is 

839 equivalent to the preceding statement:: 

840 

841 stmt = select(users_table).where( 

842 case({"wendy": "W", "jack": "J"}, value=users_table.c.name, else_="E") 

843 ) 

844 

845 The values which are accepted as result values in 

846 :paramref:`.case.whens` as well as with :paramref:`.case.else_` are 

847 coerced from Python literals into :func:`.bindparam` constructs. 

848 SQL expressions, e.g. :class:`_expression.ColumnElement` constructs, 

849 are accepted 

850 as well. To coerce a literal string expression into a constant 

851 expression rendered inline, use the :func:`_expression.literal_column` 

852 construct, 

853 as in:: 

854 

855 from sqlalchemy import case, literal_column 

856 

857 case( 

858 (orderline.c.qty > 100, literal_column("'greaterthan100'")), 

859 (orderline.c.qty > 10, literal_column("'greaterthan10'")), 

860 else_=literal_column("'lessthan10'"), 

861 ) 

862 

863 The above will render the given constants without using bound 

864 parameters for the result values (but still for the comparison 

865 values), as in: 

866 

867 .. sourcecode:: sql 

868 

869 CASE 

870 WHEN (orderline.qty > :qty_1) THEN 'greaterthan100' 

871 WHEN (orderline.qty > :qty_2) THEN 'greaterthan10' 

872 ELSE 'lessthan10' 

873 END 

874 

875 :param \*whens: The criteria to be compared against, 

876 :paramref:`.case.whens` accepts two different forms, based on 

877 whether or not :paramref:`.case.value` is used. 

878 

879 .. versionchanged:: 1.4 the :func:`_sql.case` 

880 function now accepts the series of WHEN conditions positionally 

881 

882 In the first form, it accepts multiple 2-tuples passed as positional 

883 arguments; each 2-tuple consists of ``(<sql expression>, <value>)``, 

884 where the SQL expression is a boolean expression and "value" is a 

885 resulting value, e.g.:: 

886 

887 case( 

888 (users_table.c.name == "wendy", "W"), 

889 (users_table.c.name == "jack", "J"), 

890 ) 

891 

892 In the second form, it accepts a Python dictionary of comparison 

893 values mapped to a resulting value; this form requires 

894 :paramref:`.case.value` to be present, and values will be compared 

895 using the ``==`` operator, e.g.:: 

896 

897 case({"wendy": "W", "jack": "J"}, value=users_table.c.name) 

898 

899 :param value: An optional SQL expression which will be used as a 

900 fixed "comparison point" for candidate values within a dictionary 

901 passed to :paramref:`.case.whens`. 

902 

903 :param else\_: An optional SQL expression which will be the evaluated 

904 result of the ``CASE`` construct if all expressions within 

905 :paramref:`.case.whens` evaluate to false. When omitted, most 

906 databases will produce a result of NULL if none of the "when" 

907 expressions evaluate to true. 

908 

909 

910 """ # noqa: E501 

911 return Case(*whens, value=value, else_=else_) 

912 

913 

914def cast( 

915 expression: _ColumnExpressionOrLiteralArgument[Any], 

916 type_: _TypeEngineArgument[_T], 

917) -> Cast[_T]: 

918 r"""Produce a ``CAST`` expression. 

919 

920 :func:`.cast` returns an instance of :class:`.Cast`. 

921 

922 E.g.:: 

923 

924 from sqlalchemy import cast, Numeric 

925 

926 stmt = select(cast(product_table.c.unit_price, Numeric(10, 4))) 

927 

928 The above statement will produce SQL resembling: 

929 

930 .. sourcecode:: sql 

931 

932 SELECT CAST(unit_price AS NUMERIC(10, 4)) FROM product 

933 

934 The :func:`.cast` function performs two distinct functions when 

935 used. The first is that it renders the ``CAST`` expression within 

936 the resulting SQL string. The second is that it associates the given 

937 type (e.g. :class:`.TypeEngine` class or instance) with the column 

938 expression on the Python side, which means the expression will take 

939 on the expression operator behavior associated with that type, 

940 as well as the bound-value handling and result-row-handling behavior 

941 of the type. 

942 

943 An alternative to :func:`.cast` is the :func:`.type_coerce` function. 

944 This function performs the second task of associating an expression 

945 with a specific type, but does not render the ``CAST`` expression 

946 in SQL. 

947 

948 :param expression: A SQL expression, such as a 

949 :class:`_expression.ColumnElement` 

950 expression or a Python string which will be coerced into a bound 

951 literal value. 

952 

953 :param type\_: A :class:`.TypeEngine` class or instance indicating 

954 the type to which the ``CAST`` should apply. 

955 

956 .. seealso:: 

957 

958 :ref:`tutorial_casts` 

959 

960 :func:`.try_cast` - an alternative to CAST that results in 

961 NULLs when the cast fails, instead of raising an error. 

962 Only supported by some dialects. 

963 

964 :func:`.type_coerce` - an alternative to CAST that coerces the type 

965 on the Python side only, which is often sufficient to generate the 

966 correct SQL and data coercion. 

967 

968 

969 """ 

970 return Cast(expression, type_) 

971 

972 

973def try_cast( 

974 expression: _ColumnExpressionOrLiteralArgument[Any], 

975 type_: _TypeEngineArgument[_T], 

976) -> TryCast[_T]: 

977 """Produce a ``TRY_CAST`` expression for backends which support it; 

978 this is a ``CAST`` which returns NULL for un-castable conversions. 

979 

980 In SQLAlchemy, this construct is supported **only** by the SQL Server 

981 dialect, and will raise a :class:`.CompileError` if used on other 

982 included backends. However, third party backends may also support 

983 this construct. 

984 

985 .. tip:: As :func:`_sql.try_cast` originates from the SQL Server dialect, 

986 it's importable both from ``sqlalchemy.`` as well as from 

987 ``sqlalchemy.dialects.mssql``. 

988 

989 :func:`_sql.try_cast` returns an instance of :class:`.TryCast` and 

990 generally behaves similarly to the :class:`.Cast` construct; 

991 at the SQL level, the difference between ``CAST`` and ``TRY_CAST`` 

992 is that ``TRY_CAST`` returns NULL for an un-castable expression, 

993 such as attempting to cast a string ``"hi"`` to an integer value. 

994 

995 E.g.:: 

996 

997 from sqlalchemy import select, try_cast, Numeric 

998 

999 stmt = select(try_cast(product_table.c.unit_price, Numeric(10, 4))) 

1000 

1001 The above would render on Microsoft SQL Server as: 

1002 

1003 .. sourcecode:: sql 

1004 

1005 SELECT TRY_CAST (product_table.unit_price AS NUMERIC(10, 4)) 

1006 FROM product_table 

1007 

1008 .. versionadded:: 2.0.14 :func:`.try_cast` has been 

1009 generalized from the SQL Server dialect into a general use 

1010 construct that may be supported by additional dialects. 

1011 

1012 """ 

1013 return TryCast(expression, type_) 

1014 

1015 

1016def column( 

1017 text: str, 

1018 type_: Optional[_TypeEngineArgument[_T]] = None, 

1019 is_literal: bool = False, 

1020 _selectable: Optional[FromClause] = None, 

1021) -> ColumnClause[_T]: 

1022 """Produce a :class:`.ColumnClause` object. 

1023 

1024 The :class:`.ColumnClause` is a lightweight analogue to the 

1025 :class:`_schema.Column` class. The :func:`_expression.column` 

1026 function can 

1027 be invoked with just a name alone, as in:: 

1028 

1029 from sqlalchemy import column 

1030 

1031 id, name = column("id"), column("name") 

1032 stmt = select(id, name).select_from("user") 

1033 

1034 The above statement would produce SQL like: 

1035 

1036 .. sourcecode:: sql 

1037 

1038 SELECT id, name FROM user 

1039 

1040 Once constructed, :func:`_expression.column` 

1041 may be used like any other SQL 

1042 expression element such as within :func:`_expression.select` 

1043 constructs:: 

1044 

1045 from sqlalchemy.sql import column 

1046 

1047 id, name = column("id"), column("name") 

1048 stmt = select(id, name).select_from("user") 

1049 

1050 The text handled by :func:`_expression.column` 

1051 is assumed to be handled 

1052 like the name of a database column; if the string contains mixed case, 

1053 special characters, or matches a known reserved word on the target 

1054 backend, the column expression will render using the quoting 

1055 behavior determined by the backend. To produce a textual SQL 

1056 expression that is rendered exactly without any quoting, 

1057 use :func:`_expression.literal_column` instead, 

1058 or pass ``True`` as the 

1059 value of :paramref:`_expression.column.is_literal`. Additionally, 

1060 full SQL 

1061 statements are best handled using the :func:`_expression.text` 

1062 construct. 

1063 

1064 :func:`_expression.column` can be used in a table-like 

1065 fashion by combining it with the :func:`.table` function 

1066 (which is the lightweight analogue to :class:`_schema.Table` 

1067 ) to produce 

1068 a working table construct with minimal boilerplate:: 

1069 

1070 from sqlalchemy import table, column, select 

1071 

1072 user = table( 

1073 "user", 

1074 column("id"), 

1075 column("name"), 

1076 column("description"), 

1077 ) 

1078 

1079 stmt = select(user.c.description).where(user.c.name == "wendy") 

1080 

1081 A :func:`_expression.column` / :func:`.table` 

1082 construct like that illustrated 

1083 above can be created in an 

1084 ad-hoc fashion and is not associated with any 

1085 :class:`_schema.MetaData`, DDL, or events, unlike its 

1086 :class:`_schema.Table` counterpart. 

1087 

1088 :param text: the text of the element. 

1089 

1090 :param type: :class:`_types.TypeEngine` object which can associate 

1091 this :class:`.ColumnClause` with a type. 

1092 

1093 :param is_literal: if True, the :class:`.ColumnClause` is assumed to 

1094 be an exact expression that will be delivered to the output with no 

1095 quoting rules applied regardless of case sensitive settings. the 

1096 :func:`_expression.literal_column()` function essentially invokes 

1097 :func:`_expression.column` while passing ``is_literal=True``. 

1098 

1099 .. seealso:: 

1100 

1101 :class:`_schema.Column` 

1102 

1103 :func:`_expression.literal_column` 

1104 

1105 :func:`.table` 

1106 

1107 :func:`_expression.text` 

1108 

1109 :ref:`tutorial_select_arbitrary_text` 

1110 

1111 """ 

1112 return ColumnClause(text, type_, is_literal, _selectable) 

1113 

1114 

1115@overload 

1116def desc( 

1117 column: Union[str, "ColumnElement[_T]"], 

1118) -> UnaryExpression[_T]: ... 

1119 

1120 

1121@overload 

1122def desc( 

1123 column: _ColumnExpressionOrStrLabelArgument[_T], 

1124) -> Union[OrderByList, UnaryExpression[_T]]: ... 

1125 

1126 

1127def desc( 

1128 column: _ColumnExpressionOrStrLabelArgument[_T], 

1129) -> Union[OrderByList, UnaryExpression[_T]]: 

1130 """Produce a descending ``ORDER BY`` clause element. 

1131 

1132 e.g.:: 

1133 

1134 from sqlalchemy import desc 

1135 

1136 stmt = select(users_table).order_by(desc(users_table.c.name)) 

1137 

1138 will produce SQL as: 

1139 

1140 .. sourcecode:: sql 

1141 

1142 SELECT id, name FROM user ORDER BY name DESC 

1143 

1144 The :func:`.desc` function is a standalone version of the 

1145 :meth:`_expression.ColumnElement.desc` 

1146 method available on all SQL expressions, 

1147 e.g.:: 

1148 

1149 

1150 stmt = select(users_table).order_by(users_table.c.name.desc()) 

1151 

1152 :param column: A :class:`_expression.ColumnElement` (e.g. 

1153 scalar SQL expression) 

1154 with which to apply the :func:`.desc` operation. 

1155 

1156 .. seealso:: 

1157 

1158 :func:`.asc` 

1159 

1160 :func:`.nulls_first` 

1161 

1162 :func:`.nulls_last` 

1163 

1164 :meth:`_expression.Select.order_by` 

1165 

1166 """ 

1167 if isinstance(column, operators.OrderingOperators): 

1168 return column.desc() # type: ignore[unused-ignore] 

1169 else: 

1170 return UnaryExpression._create_desc(column) 

1171 

1172 

1173def distinct(expr: _ColumnExpressionArgument[_T]) -> UnaryExpression[_T]: 

1174 """Produce an column-expression-level unary ``DISTINCT`` clause. 

1175 

1176 This applies the ``DISTINCT`` keyword to an **individual column 

1177 expression** (e.g. not the whole statement), and renders **specifically 

1178 in that column position**; this is used for containment within 

1179 an aggregate function, as in:: 

1180 

1181 from sqlalchemy import distinct, func 

1182 

1183 stmt = select(users_table.c.id, func.count(distinct(users_table.c.name))) 

1184 

1185 The above would produce an statement resembling: 

1186 

1187 .. sourcecode:: sql 

1188 

1189 SELECT user.id, count(DISTINCT user.name) FROM user 

1190 

1191 .. tip:: The :func:`_sql.distinct` function does **not** apply DISTINCT 

1192 to the full SELECT statement, instead applying a DISTINCT modifier 

1193 to **individual column expressions**. For general ``SELECT DISTINCT`` 

1194 support, use the 

1195 :meth:`_sql.Select.distinct` method on :class:`_sql.Select`. 

1196 

1197 The :func:`.distinct` function is also available as a column-level 

1198 method, e.g. :meth:`_expression.ColumnElement.distinct`, as in:: 

1199 

1200 stmt = select(func.count(users_table.c.name.distinct())) 

1201 

1202 The :func:`.distinct` operator is different from the 

1203 :meth:`_expression.Select.distinct` method of 

1204 :class:`_expression.Select`, 

1205 which produces a ``SELECT`` statement 

1206 with ``DISTINCT`` applied to the result set as a whole, 

1207 e.g. a ``SELECT DISTINCT`` expression. See that method for further 

1208 information. 

1209 

1210 .. seealso:: 

1211 

1212 :meth:`_expression.ColumnElement.distinct` 

1213 

1214 :meth:`_expression.Select.distinct` 

1215 

1216 :data:`.func` 

1217 

1218 """ # noqa: E501 

1219 if isinstance(expr, operators.ColumnOperators): 

1220 return expr.distinct() 

1221 else: 

1222 return UnaryExpression._create_distinct(expr) 

1223 

1224 

1225def bitwise_not(expr: _ColumnExpressionArgument[_T]) -> UnaryExpression[_T]: 

1226 """Produce a unary bitwise NOT clause, typically via the ``~`` operator. 

1227 

1228 Not to be confused with boolean negation :func:`_sql.not_`. 

1229 

1230 .. versionadded:: 2.0.2 

1231 

1232 .. seealso:: 

1233 

1234 :ref:`operators_bitwise` 

1235 

1236 

1237 """ 

1238 if isinstance(expr, operators.ColumnOperators): 

1239 return expr.bitwise_not() 

1240 else: 

1241 return UnaryExpression._create_bitwise_not(expr) 

1242 

1243 

1244def extract(field: str, expr: _ColumnExpressionArgument[Any]) -> Extract: 

1245 """Return a :class:`.Extract` construct. 

1246 

1247 This is typically available as :func:`.extract` 

1248 as well as ``func.extract`` from the 

1249 :data:`.func` namespace. 

1250 

1251 :param field: The field to extract. 

1252 

1253 .. warning:: This field is used as a literal SQL string. 

1254 **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**. 

1255 

1256 :param expr: A column or Python scalar expression serving as the 

1257 right side of the ``EXTRACT`` expression. 

1258 

1259 E.g.:: 

1260 

1261 from sqlalchemy import extract 

1262 from sqlalchemy import table, column 

1263 

1264 logged_table = table( 

1265 "user", 

1266 column("id"), 

1267 column("date_created"), 

1268 ) 

1269 

1270 stmt = select(logged_table.c.id).where( 

1271 extract("YEAR", logged_table.c.date_created) == 2021 

1272 ) 

1273 

1274 In the above example, the statement is used to select ids from the 

1275 database where the ``YEAR`` component matches a specific value. 

1276 

1277 Similarly, one can also select an extracted component:: 

1278 

1279 stmt = select(extract("YEAR", logged_table.c.date_created)).where( 

1280 logged_table.c.id == 1 

1281 ) 

1282 

1283 The implementation of ``EXTRACT`` may vary across database backends. 

1284 Users are reminded to consult their database documentation. 

1285 """ 

1286 return Extract(field, expr) 

1287 

1288 

1289def false() -> False_: 

1290 """Return a :class:`.False_` construct. 

1291 

1292 E.g.: 

1293 

1294 .. sourcecode:: pycon+sql 

1295 

1296 >>> from sqlalchemy import false 

1297 >>> print(select(t.c.x).where(false())) 

1298 {printsql}SELECT x FROM t WHERE false 

1299 

1300 A backend which does not support true/false constants will render as 

1301 an expression against 1 or 0: 

1302 

1303 .. sourcecode:: pycon+sql 

1304 

1305 >>> print(select(t.c.x).where(false())) 

1306 {printsql}SELECT x FROM t WHERE 0 = 1 

1307 

1308 The :func:`.true` and :func:`.false` constants also feature 

1309 "short circuit" operation within an :func:`.and_` or :func:`.or_` 

1310 conjunction: 

1311 

1312 .. sourcecode:: pycon+sql 

1313 

1314 >>> print(select(t.c.x).where(or_(t.c.x > 5, true()))) 

1315 {printsql}SELECT x FROM t WHERE true{stop} 

1316 

1317 >>> print(select(t.c.x).where(and_(t.c.x > 5, false()))) 

1318 {printsql}SELECT x FROM t WHERE false{stop} 

1319 

1320 .. seealso:: 

1321 

1322 :func:`.true` 

1323 

1324 """ 

1325 

1326 return False_._instance() 

1327 

1328 

1329def funcfilter( 

1330 func: FunctionElement[_T], *criterion: _ColumnExpressionArgument[bool] 

1331) -> FunctionFilter[_T]: 

1332 """Produce a :class:`.FunctionFilter` object against a function. 

1333 

1334 Used against aggregate and window functions, 

1335 for database backends that support the "FILTER" clause. 

1336 

1337 E.g.:: 

1338 

1339 from sqlalchemy import funcfilter 

1340 

1341 funcfilter(func.count(1), MyClass.name == "some name") 

1342 

1343 Would produce "COUNT(1) FILTER (WHERE myclass.name = 'some name')". 

1344 

1345 This function is also available from the :data:`~.expression.func` 

1346 construct itself via the :meth:`.FunctionElement.filter` method. 

1347 

1348 .. seealso:: 

1349 

1350 :ref:`tutorial_functions_within_group` - in the 

1351 :ref:`unified_tutorial` 

1352 

1353 :meth:`.FunctionElement.filter` 

1354 

1355 """ 

1356 return FunctionFilter(func, *criterion) 

1357 

1358 

1359def label( 

1360 name: str, 

1361 element: _ColumnExpressionArgument[_T], 

1362 type_: Optional[_TypeEngineArgument[_T]] = None, 

1363) -> Label[_T]: 

1364 """Return a :class:`Label` object for the 

1365 given :class:`_expression.ColumnElement`. 

1366 

1367 A label changes the name of an element in the columns clause of a 

1368 ``SELECT`` statement, typically via the ``AS`` SQL keyword. 

1369 

1370 This functionality is more conveniently available via the 

1371 :meth:`_expression.ColumnElement.label` method on 

1372 :class:`_expression.ColumnElement`. 

1373 

1374 :param name: label name 

1375 

1376 :param obj: a :class:`_expression.ColumnElement`. 

1377 

1378 """ 

1379 return Label(name, element, type_) 

1380 

1381 

1382def null() -> Null: 

1383 """Return a constant :class:`.Null` construct.""" 

1384 

1385 return Null._instance() 

1386 

1387 

1388@overload 

1389def nulls_first( 

1390 column: "ColumnElement[_T]", 

1391) -> UnaryExpression[_T]: ... 

1392 

1393 

1394@overload 

1395def nulls_first( 

1396 column: _ColumnExpressionArgument[_T], 

1397) -> Union[OrderByList, UnaryExpression[_T]]: ... 

1398 

1399 

1400def nulls_first( 

1401 column: _ColumnExpressionArgument[_T], 

1402) -> Union[OrderByList, UnaryExpression[_T]]: 

1403 """Produce the ``NULLS FIRST`` modifier for an ``ORDER BY`` expression. 

1404 

1405 :func:`.nulls_first` is intended to modify the expression produced 

1406 by :func:`.asc` or :func:`.desc`, and indicates how NULL values 

1407 should be handled when they are encountered during ordering:: 

1408 

1409 

1410 from sqlalchemy import desc, nulls_first 

1411 

1412 stmt = select(users_table).order_by(nulls_first(desc(users_table.c.name))) 

1413 

1414 The SQL expression from the above would resemble: 

1415 

1416 .. sourcecode:: sql 

1417 

1418 SELECT id, name FROM user ORDER BY name DESC NULLS FIRST 

1419 

1420 Like :func:`.asc` and :func:`.desc`, :func:`.nulls_first` is typically 

1421 invoked from the column expression itself using 

1422 :meth:`_expression.ColumnElement.nulls_first`, 

1423 rather than as its standalone 

1424 function version, as in:: 

1425 

1426 stmt = select(users_table).order_by( 

1427 users_table.c.name.desc().nulls_first() 

1428 ) 

1429 

1430 .. versionchanged:: 1.4 :func:`.nulls_first` is renamed from 

1431 :func:`.nullsfirst` in previous releases. 

1432 The previous name remains available for backwards compatibility. 

1433 

1434 .. seealso:: 

1435 

1436 :func:`.asc` 

1437 

1438 :func:`.desc` 

1439 

1440 :func:`.nulls_last` 

1441 

1442 :meth:`_expression.Select.order_by` 

1443 

1444 """ # noqa: E501 

1445 if isinstance(column, operators.OrderingOperators): 

1446 return column.nulls_first() 

1447 else: 

1448 return UnaryExpression._create_nulls_first(column) 

1449 

1450 

1451@overload 

1452def nulls_last( 

1453 column: "ColumnElement[_T]", 

1454) -> UnaryExpression[_T]: ... 

1455 

1456 

1457@overload 

1458def nulls_last( 

1459 column: _ColumnExpressionArgument[_T], 

1460) -> Union[OrderByList, UnaryExpression[_T]]: ... 

1461 

1462 

1463def nulls_last( 

1464 column: _ColumnExpressionArgument[_T], 

1465) -> Union[OrderByList, UnaryExpression[_T]]: 

1466 """Produce the ``NULLS LAST`` modifier for an ``ORDER BY`` expression. 

1467 

1468 :func:`.nulls_last` is intended to modify the expression produced 

1469 by :func:`.asc` or :func:`.desc`, and indicates how NULL values 

1470 should be handled when they are encountered during ordering:: 

1471 

1472 

1473 from sqlalchemy import desc, nulls_last 

1474 

1475 stmt = select(users_table).order_by(nulls_last(desc(users_table.c.name))) 

1476 

1477 The SQL expression from the above would resemble: 

1478 

1479 .. sourcecode:: sql 

1480 

1481 SELECT id, name FROM user ORDER BY name DESC NULLS LAST 

1482 

1483 Like :func:`.asc` and :func:`.desc`, :func:`.nulls_last` is typically 

1484 invoked from the column expression itself using 

1485 :meth:`_expression.ColumnElement.nulls_last`, 

1486 rather than as its standalone 

1487 function version, as in:: 

1488 

1489 stmt = select(users_table).order_by(users_table.c.name.desc().nulls_last()) 

1490 

1491 .. versionchanged:: 1.4 :func:`.nulls_last` is renamed from 

1492 :func:`.nullslast` in previous releases. 

1493 The previous name remains available for backwards compatibility. 

1494 

1495 .. seealso:: 

1496 

1497 :func:`.asc` 

1498 

1499 :func:`.desc` 

1500 

1501 :func:`.nulls_first` 

1502 

1503 :meth:`_expression.Select.order_by` 

1504 

1505 """ # noqa: E501 

1506 if isinstance(column, operators.OrderingOperators): 

1507 return column.nulls_last() 

1508 else: 

1509 return UnaryExpression._create_nulls_last(column) 

1510 

1511 

1512def or_( # type: ignore[empty-body] 

1513 initial_clause: Union[Literal[False], _ColumnExpressionArgument[bool]], 

1514 *clauses: _ColumnExpressionArgument[bool], 

1515) -> ColumnElement[bool]: 

1516 """Produce a conjunction of expressions joined by ``OR``. 

1517 

1518 E.g.:: 

1519 

1520 from sqlalchemy import or_ 

1521 

1522 stmt = select(users_table).where( 

1523 or_(users_table.c.name == "wendy", users_table.c.name == "jack") 

1524 ) 

1525 

1526 The :func:`.or_` conjunction is also available using the 

1527 Python ``|`` operator (though note that compound expressions 

1528 need to be parenthesized in order to function with Python 

1529 operator precedence behavior):: 

1530 

1531 stmt = select(users_table).where( 

1532 (users_table.c.name == "wendy") | (users_table.c.name == "jack") 

1533 ) 

1534 

1535 The :func:`.or_` construct must be given at least one positional 

1536 argument in order to be valid; a :func:`.or_` construct with no 

1537 arguments is ambiguous. To produce an "empty" or dynamically 

1538 generated :func:`.or_` expression, from a given list of expressions, 

1539 a "default" element of :func:`_sql.false` (or just ``False``) should be 

1540 specified:: 

1541 

1542 from sqlalchemy import false 

1543 

1544 or_criteria = or_(false(), *expressions) 

1545 

1546 The above expression will compile to SQL as the expression ``false`` 

1547 or ``0 = 1``, depending on backend, if no other expressions are 

1548 present. If expressions are present, then the :func:`_sql.false` value is 

1549 ignored as it does not affect the outcome of an OR expression which 

1550 has other elements. 

1551 

1552 .. deprecated:: 1.4 The :func:`.or_` element now requires that at 

1553 least one argument is passed; creating the :func:`.or_` construct 

1554 with no arguments is deprecated, and will emit a deprecation warning 

1555 while continuing to produce a blank SQL string. 

1556 

1557 .. seealso:: 

1558 

1559 :func:`.and_` 

1560 

1561 """ 

1562 ... 

1563 

1564 

1565if not TYPE_CHECKING: 

1566 # handle deprecated case which allows zero-arguments 

1567 def or_(*clauses): # noqa: F811 

1568 """Produce a conjunction of expressions joined by ``OR``. 

1569 

1570 E.g.:: 

1571 

1572 from sqlalchemy import or_ 

1573 

1574 stmt = select(users_table).where( 

1575 or_(users_table.c.name == "wendy", users_table.c.name == "jack") 

1576 ) 

1577 

1578 The :func:`.or_` conjunction is also available using the 

1579 Python ``|`` operator (though note that compound expressions 

1580 need to be parenthesized in order to function with Python 

1581 operator precedence behavior):: 

1582 

1583 stmt = select(users_table).where( 

1584 (users_table.c.name == "wendy") | (users_table.c.name == "jack") 

1585 ) 

1586 

1587 The :func:`.or_` construct must be given at least one positional 

1588 argument in order to be valid; a :func:`.or_` construct with no 

1589 arguments is ambiguous. To produce an "empty" or dynamically 

1590 generated :func:`.or_` expression, from a given list of expressions, 

1591 a "default" element of :func:`_sql.false` (or just ``False``) should be 

1592 specified:: 

1593 

1594 from sqlalchemy import false 

1595 

1596 or_criteria = or_(false(), *expressions) 

1597 

1598 The above expression will compile to SQL as the expression ``false`` 

1599 or ``0 = 1``, depending on backend, if no other expressions are 

1600 present. If expressions are present, then the :func:`_sql.false` value 

1601 is ignored as it does not affect the outcome of an OR expression which 

1602 has other elements. 

1603 

1604 .. deprecated:: 1.4 The :func:`.or_` element now requires that at 

1605 least one argument is passed; creating the :func:`.or_` construct 

1606 with no arguments is deprecated, and will emit a deprecation warning 

1607 while continuing to produce a blank SQL string. 

1608 

1609 .. seealso:: 

1610 

1611 :func:`.and_` 

1612 

1613 """ # noqa: E501 

1614 return BooleanClauseList.or_(*clauses) 

1615 

1616 

1617def over( 

1618 element: FunctionElement[_T], 

1619 partition_by: _ByArgument | None = None, 

1620 order_by: _ByArgument | None = None, 

1621 range_: _FrameIntTuple | FrameClause | None = None, 

1622 rows: _FrameIntTuple | FrameClause | None = None, 

1623 groups: _FrameIntTuple | FrameClause | None = None, 

1624) -> Over[_T]: 

1625 r"""Produce an :class:`.Over` object against a function. 

1626 

1627 Used against aggregate or so-called "window" functions, 

1628 for database backends that support window functions. 

1629 

1630 :func:`_expression.over` is usually called using 

1631 the :meth:`.FunctionElement.over` method, e.g.:: 

1632 

1633 func.row_number().over(order_by=mytable.c.some_column) 

1634 

1635 Would produce: 

1636 

1637 .. sourcecode:: sql 

1638 

1639 ROW_NUMBER() OVER(ORDER BY some_column) 

1640 

1641 Ranges are also possible using the :paramref:`.expression.over.range_`, 

1642 :paramref:`.expression.over.rows`, and :paramref:`.expression.over.groups` 

1643 parameters. These 

1644 mutually-exclusive parameters each accept a 2-tuple, which contains 

1645 a combination of integers and None:: 

1646 

1647 func.row_number().over(order_by=my_table.c.some_column, range_=(None, 0)) 

1648 

1649 The above would produce: 

1650 

1651 .. sourcecode:: sql 

1652 

1653 ROW_NUMBER() OVER(ORDER BY some_column 

1654 RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) 

1655 

1656 A value of ``None`` indicates "unbounded", a 

1657 value of zero indicates "current row", and negative / positive 

1658 integers indicate "preceding" and "following": 

1659 

1660 * RANGE BETWEEN 5 PRECEDING AND 10 FOLLOWING:: 

1661 

1662 func.row_number().over(order_by="x", range_=(-5, 10)) 

1663 

1664 * ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW:: 

1665 

1666 func.row_number().over(order_by="x", rows=(None, 0)) 

1667 

1668 * RANGE BETWEEN 2 PRECEDING AND UNBOUNDED FOLLOWING:: 

1669 

1670 func.row_number().over(order_by="x", range_=(-2, None)) 

1671 

1672 * RANGE BETWEEN 1 FOLLOWING AND 3 FOLLOWING:: 

1673 

1674 func.row_number().over(order_by="x", range_=(1, 3)) 

1675 

1676 * GROUPS BETWEEN 1 FOLLOWING AND 3 FOLLOWING:: 

1677 

1678 func.row_number().over(order_by="x", groups=(1, 3)) 

1679 

1680 Depending on the type of the order column, the 'RANGE' value may not be 

1681 an integer. In this case use a :class:`_expression.FrameClause` directly 

1682 to specify the frame boundaries. E.g.:: 

1683 

1684 from datetime import timedelta 

1685 from sqlalchemy import FrameClause, FrameClauseType 

1686 

1687 func.sum(my_table.c.amount).over( 

1688 order_by=my_table.c.date, 

1689 range_=FrameClause( 

1690 start=timedelta(days=7), 

1691 end=None, 

1692 start_frame_type=FrameClauseType.PRECEDING, 

1693 end_frame_type=FrameClauseType.UNBOUNDED, 

1694 ), 

1695 ) 

1696 

1697 .. versionchanged:: 2.1 Added support for range types that are not 

1698 integer-based, via the :class:`_expression.FrameClause` construct. 

1699 

1700 :param element: a :class:`.FunctionElement`, :class:`.WithinGroup`, 

1701 or other compatible construct. 

1702 :param partition_by: a column element or string, or a list 

1703 of such, that will be used as the PARTITION BY clause 

1704 of the OVER construct. 

1705 :param order_by: a column element or string, or a list 

1706 of such, that will be used as the ORDER BY clause 

1707 of the OVER construct. 

1708 :param range\_: optional range clause for the window. This is a 

1709 two-tuple value which can contain integer values or ``None``, 

1710 and will render a RANGE BETWEEN PRECEDING / FOLLOWING clause. 

1711 Can also be a :class:`_expression.FrameClause` instance to 

1712 specify non-integer values. 

1713 

1714 .. versionchanged:: 2.1 Added support for range types that are not 

1715 integer-based, via the :class:`_expression.FrameClause` construct. 

1716 

1717 :param rows: optional rows clause for the window. This is a two-tuple 

1718 value which can contain integer values or None, and will render 

1719 a ROWS BETWEEN PRECEDING / FOLLOWING clause. Can also be a 

1720 :class:`_expression.FrameClause` instance. 

1721 :param groups: optional groups clause for the window. This is a 

1722 two-tuple value which can contain integer values or ``None``, 

1723 and will render a GROUPS BETWEEN PRECEDING / FOLLOWING clause. 

1724 Can also be a :class:`_expression.FrameClause` instance. 

1725 

1726 .. versionadded:: 2.0.40 

1727 

1728 This function is also available from the :data:`~.expression.func` 

1729 construct itself via the :meth:`.FunctionElement.over` method. 

1730 

1731 .. seealso:: 

1732 

1733 :ref:`tutorial_window_functions` - in the :ref:`unified_tutorial` 

1734 

1735 :data:`.expression.func` 

1736 

1737 :func:`_expression.within_group` 

1738 

1739 """ # noqa: E501 

1740 return Over(element, partition_by, order_by, range_, rows, groups) 

1741 

1742 

1743@_document_text_coercion("text", ":func:`.text`", ":paramref:`.text.text`") 

1744def text(text: str) -> TextClause: 

1745 r"""Construct a new :class:`_expression.TextClause` clause, 

1746 representing 

1747 a textual SQL string directly. 

1748 

1749 E.g.:: 

1750 

1751 from sqlalchemy import text 

1752 

1753 t = text("SELECT * FROM users") 

1754 result = connection.execute(t) 

1755 

1756 The advantages :func:`_expression.text` 

1757 provides over a plain string are 

1758 backend-neutral support for bind parameters, per-statement 

1759 execution options, as well as 

1760 bind parameter and result-column typing behavior, allowing 

1761 SQLAlchemy type constructs to play a role when executing 

1762 a statement that is specified literally. The construct can also 

1763 be provided with a ``.c`` collection of column elements, allowing 

1764 it to be embedded in other SQL expression constructs as a subquery. 

1765 

1766 Bind parameters are specified by name, using the format ``:name``. 

1767 E.g.:: 

1768 

1769 t = text("SELECT * FROM users WHERE id=:user_id") 

1770 result = connection.execute(t, {"user_id": 12}) 

1771 

1772 For SQL statements where a colon is required verbatim, as within 

1773 an inline string, use a backslash to escape:: 

1774 

1775 t = text(r"SELECT * FROM users WHERE name='\:username'") 

1776 

1777 The :class:`_expression.TextClause` 

1778 construct includes methods which can 

1779 provide information about the bound parameters as well as the column 

1780 values which would be returned from the textual statement, assuming 

1781 it's an executable SELECT type of statement. The 

1782 :meth:`_expression.TextClause.bindparams` 

1783 method is used to provide bound 

1784 parameter detail, and :meth:`_expression.TextClause.columns` 

1785 method allows 

1786 specification of return columns including names and types:: 

1787 

1788 t = ( 

1789 text("SELECT * FROM users WHERE id=:user_id") 

1790 .bindparams(user_id=7) 

1791 .columns(id=Integer, name=String) 

1792 ) 

1793 

1794 for id, name in connection.execute(t): 

1795 print(id, name) 

1796 

1797 The :func:`_expression.text` construct is used in cases when 

1798 a literal string SQL fragment is specified as part of a larger query, 

1799 such as for the WHERE clause of a SELECT statement:: 

1800 

1801 s = select(users.c.id, users.c.name).where(text("id=:user_id")) 

1802 result = connection.execute(s, {"user_id": 12}) 

1803 

1804 :func:`_expression.text` is also used for the construction 

1805 of a full, standalone statement using plain text. 

1806 As such, SQLAlchemy refers 

1807 to it as an :class:`.Executable` object and may be used 

1808 like any other statement passed to an ``.execute()`` method. 

1809 

1810 :param text: 

1811 the text of the SQL statement to be created. Use ``:<param>`` 

1812 to specify bind parameters; they will be compiled to their 

1813 engine-specific format. 

1814 

1815 .. seealso:: 

1816 

1817 :ref:`tutorial_select_arbitrary_text` 

1818 

1819 """ 

1820 return TextClause(text) 

1821 

1822 

1823def tstring(template: Template) -> TString: 

1824 r"""Construct a new :class:`_expression.TString` clause, 

1825 representing a SQL template string using Python 3.14+ t-strings. 

1826 

1827 .. versionadded:: 2.1 

1828 

1829 E.g.:: 

1830 

1831 from sqlalchemy import tstring 

1832 

1833 a = 5 

1834 b = 10 

1835 stmt = tstring(t"select {a}, {b}") 

1836 result = connection.execute(stmt) 

1837 

1838 The :func:`_expression.tstring` function accepts a Python 3.14+ 

1839 template string (t-string) and processes it to create a SQL statement. 

1840 Unlike :func:`_expression.text`, which requires manual bind parameter 

1841 specification, :func:`_expression.tstring` automatically handles 

1842 interpolation of Python values and SQLAlchemy expressions. 

1843 

1844 **Interpolation Behavior**: 

1845 

1846 - **SQL content** expressed in the plain string portions of the template 

1847 are rendered directly as SQL 

1848 - **SQLAlchemy expressions** (columns, functions, etc.) are embedded 

1849 as clause elements 

1850 - **Plain Python values** are automatically wrapped in 

1851 :func:`_expression.literal` 

1852 

1853 For example:: 

1854 

1855 from sqlalchemy import tstring, select, literal, JSON, table, column 

1856 

1857 # Python values become bound parameters 

1858 user_id = 42 

1859 stmt = tstring(t"SELECT * FROM users WHERE id = {user_id}") 

1860 # renders: SELECT * FROM users WHERE id = :param_1 

1861 

1862 # SQLAlchemy expressions are embedded 

1863 stmt = tstring(t"SELECT {column('q')} FROM {table('t')}") 

1864 # renders: SELECT q FROM t 

1865 

1866 # Apply explicit SQL types to bound values using literal() 

1867 some_json = {"foo": "bar"} 

1868 stmt = tstring(t"SELECT {literal(some_json, JSON)}") 

1869 

1870 **Column Specification**: 

1871 

1872 Like :func:`_expression.text`, the :func:`_expression.tstring` construct 

1873 supports the :meth:`_expression.TString.columns` method to specify 

1874 return columns and their types:: 

1875 

1876 from sqlalchemy import tstring, column, Integer, String 

1877 

1878 stmt = tstring(t"SELECT id, name FROM users").columns( 

1879 column("id", Integer), column("name", String) 

1880 ) 

1881 

1882 for id, name in connection.execute(stmt): 

1883 print(id, name) 

1884 

1885 :param template: 

1886 a Python 3.14+ template string (t-string) containing SQL fragments 

1887 and Python expressions to be interpolated. 

1888 

1889 .. seealso:: 

1890 

1891 :ref:`tutorial_select_arbitrary_text` - in the :ref:`unified_tutorial` 

1892 

1893 :class:`_expression.TString` 

1894 

1895 :func:`_expression.text` 

1896 

1897 `PEP 750 <https://peps.python.org/pep-0750/>`_ - Template Strings 

1898 

1899 """ 

1900 return TString(template) 

1901 

1902 

1903def true() -> True_: 

1904 """Return a constant :class:`.True_` construct. 

1905 

1906 E.g.: 

1907 

1908 .. sourcecode:: pycon+sql 

1909 

1910 >>> from sqlalchemy import true 

1911 >>> print(select(t.c.x).where(true())) 

1912 {printsql}SELECT x FROM t WHERE true 

1913 

1914 A backend which does not support true/false constants will render as 

1915 an expression against 1 or 0: 

1916 

1917 .. sourcecode:: pycon+sql 

1918 

1919 >>> print(select(t.c.x).where(true())) 

1920 {printsql}SELECT x FROM t WHERE 1 = 1 

1921 

1922 The :func:`.true` and :func:`.false` constants also feature 

1923 "short circuit" operation within an :func:`.and_` or :func:`.or_` 

1924 conjunction: 

1925 

1926 .. sourcecode:: pycon+sql 

1927 

1928 >>> print(select(t.c.x).where(or_(t.c.x > 5, true()))) 

1929 {printsql}SELECT x FROM t WHERE true{stop} 

1930 

1931 >>> print(select(t.c.x).where(and_(t.c.x > 5, false()))) 

1932 {printsql}SELECT x FROM t WHERE false{stop} 

1933 

1934 .. seealso:: 

1935 

1936 :func:`.false` 

1937 

1938 """ 

1939 

1940 return True_._instance() 

1941 

1942 

1943def tuple_( 

1944 *clauses: _ColumnExpressionOrLiteralArgument[Any], 

1945 types: Optional[Sequence[_TypeEngineArgument[Any]]] = None, 

1946) -> Tuple: 

1947 """Return a :class:`.Tuple`. 

1948 

1949 Main usage is to produce a composite IN construct using 

1950 :meth:`.ColumnOperators.in_` :: 

1951 

1952 from sqlalchemy import tuple_ 

1953 

1954 tuple_(table.c.col1, table.c.col2).in_([(1, 2), (5, 12), (10, 19)]) 

1955 

1956 .. warning:: 

1957 

1958 The composite IN construct is not supported by all backends, and is 

1959 currently known to work on PostgreSQL, MySQL, and SQLite. 

1960 Unsupported backends will raise a subclass of 

1961 :class:`~sqlalchemy.exc.DBAPIError` when such an expression is 

1962 invoked. 

1963 

1964 """ 

1965 return Tuple(*clauses, types=types) 

1966 

1967 

1968def type_coerce( 

1969 expression: _ColumnExpressionOrLiteralArgument[Any], 

1970 type_: _TypeEngineArgument[_T], 

1971) -> TypeCoerce[_T]: 

1972 r"""Associate a SQL expression with a particular type, without rendering 

1973 ``CAST``. 

1974 

1975 E.g.:: 

1976 

1977 from sqlalchemy import type_coerce 

1978 

1979 stmt = select(type_coerce(log_table.date_string, StringDateTime())) 

1980 

1981 The above construct will produce a :class:`.TypeCoerce` object, which 

1982 does not modify the rendering in any way on the SQL side, with the 

1983 possible exception of a generated label if used in a columns clause 

1984 context: 

1985 

1986 .. sourcecode:: sql 

1987 

1988 SELECT date_string AS date_string FROM log 

1989 

1990 When result rows are fetched, the ``StringDateTime`` type processor 

1991 will be applied to result rows on behalf of the ``date_string`` column. 

1992 

1993 .. note:: the :func:`.type_coerce` construct does not render any 

1994 SQL syntax of its own, including that it does not imply 

1995 parenthesization. Please use :meth:`.TypeCoerce.self_group` 

1996 if explicit parenthesization is required. 

1997 

1998 In order to provide a named label for the expression, use 

1999 :meth:`_expression.ColumnElement.label`:: 

2000 

2001 stmt = select( 

2002 type_coerce(log_table.date_string, StringDateTime()).label("date") 

2003 ) 

2004 

2005 A type that features bound-value handling will also have that behavior 

2006 take effect when literal values or :func:`.bindparam` constructs are 

2007 passed to :func:`.type_coerce` as targets. 

2008 For example, if a type implements the 

2009 :meth:`.TypeEngine.bind_expression` 

2010 method or :meth:`.TypeEngine.bind_processor` method or equivalent, 

2011 these functions will take effect at statement compilation/execution 

2012 time when a literal value is passed, as in:: 

2013 

2014 # bound-value handling of MyStringType will be applied to the 

2015 # literal value "some string" 

2016 stmt = select(type_coerce("some string", MyStringType)) 

2017 

2018 When using :func:`.type_coerce` with composed expressions, note that 

2019 **parenthesis are not applied**. If :func:`.type_coerce` is being 

2020 used in an operator context where the parenthesis normally present from 

2021 CAST are necessary, use the :meth:`.TypeCoerce.self_group` method: 

2022 

2023 .. sourcecode:: pycon+sql 

2024 

2025 >>> some_integer = column("someint", Integer) 

2026 >>> some_string = column("somestr", String) 

2027 >>> expr = type_coerce(some_integer + 5, String) + some_string 

2028 >>> print(expr) 

2029 {printsql}someint + :someint_1 || somestr{stop} 

2030 >>> expr = type_coerce(some_integer + 5, String).self_group() + some_string 

2031 >>> print(expr) 

2032 {printsql}(someint + :someint_1) || somestr{stop} 

2033 

2034 :param expression: A SQL expression, such as a 

2035 :class:`_expression.ColumnElement` 

2036 expression or a Python string which will be coerced into a bound 

2037 literal value. 

2038 

2039 :param type\_: A :class:`.TypeEngine` class or instance indicating 

2040 the type to which the expression is coerced. 

2041 

2042 .. seealso:: 

2043 

2044 :ref:`tutorial_casts` 

2045 

2046 :func:`.cast` 

2047 

2048 """ # noqa 

2049 return TypeCoerce(expression, type_) 

2050 

2051 

2052def within_group( 

2053 element: FunctionElement[_T], *order_by: _ColumnExpressionArgument[Any] 

2054) -> WithinGroup[_T]: 

2055 r"""Produce a :class:`.WithinGroup` object against a function. 

2056 

2057 Used against so-called "ordered set aggregate" and "hypothetical 

2058 set aggregate" functions, including :class:`.percentile_cont`, 

2059 :class:`.rank`, :class:`.dense_rank`, etc. This feature is typically 

2060 used by Oracle Database, Microsoft SQL Server. 

2061 

2062 For generalized ORDER BY of aggregate functions on all included 

2063 backends, including PostgreSQL, MySQL/MariaDB, SQLite as well as Oracle 

2064 and SQL Server, the :func:`_sql.aggregate_order_by` provides a more 

2065 general approach that compiles to "WITHIN GROUP" only on those backends 

2066 which require it. 

2067 

2068 :func:`_expression.within_group` is usually called using 

2069 the :meth:`.FunctionElement.within_group` method, e.g.:: 

2070 

2071 stmt = select( 

2072 func.percentile_cont(0.5).within_group(department.c.salary.desc()), 

2073 ) 

2074 

2075 The above statement would produce SQL similar to 

2076 ``SELECT percentile_cont(0.5) 

2077 WITHIN GROUP (ORDER BY department.salary DESC)``. 

2078 

2079 :param element: a :class:`.FunctionElement` construct, typically 

2080 generated by :data:`~.expression.func`. 

2081 :param \*order_by: one or more column elements that will be used 

2082 as the ORDER BY clause of the WITHIN GROUP construct. 

2083 

2084 .. seealso:: 

2085 

2086 :ref:`tutorial_functions_within_group` - in the 

2087 :ref:`unified_tutorial` 

2088 

2089 :func:`_sql.aggregate_order_by` - helper for PostgreSQL, MySQL, 

2090 SQLite aggregate functions 

2091 

2092 :data:`.expression.func` 

2093 

2094 :func:`_expression.over` 

2095 

2096 """ 

2097 return WithinGroup(element, *order_by) 

2098 

2099 

2100def aggregate_order_by( 

2101 element: FunctionElement[_T], *order_by: _ColumnExpressionArgument[Any] 

2102) -> AggregateOrderBy[_T]: 

2103 r"""Produce a :class:`.AggregateOrderBy` object against a function. 

2104 

2105 Used for aggregating functions such as :class:`_functions.array_agg`, 

2106 ``group_concat``, ``json_agg`` on backends that support ordering via an 

2107 embedded ``ORDER BY`` parameter, e.g. PostgreSQL, MySQL/MariaDB, SQLite. 

2108 When used on backends like Oracle and SQL Server, SQL compilation uses that 

2109 of :class:`.WithinGroup`. On PostgreSQL, compilation is fixed at embedded 

2110 ``ORDER BY``; for set aggregation functions where PostgreSQL requires the 

2111 use of ``WITHIN GROUP``, :func:`_expression.within_group` should be used 

2112 explicitly. 

2113 

2114 :func:`_expression.aggregate_order_by` is usually called using 

2115 the :meth:`.FunctionElement.aggregate_order_by` method, e.g.:: 

2116 

2117 stmt = select( 

2118 func.array_agg(department.c.code).aggregate_order_by( 

2119 department.c.code.desc() 

2120 ), 

2121 ) 

2122 

2123 which would produce an expression resembling: 

2124 

2125 .. sourcecode:: sql 

2126 

2127 SELECT array_agg(department.code ORDER BY department.code DESC) 

2128 AS array_agg_1 FROM department 

2129 

2130 The ORDER BY argument may also be multiple terms. 

2131 

2132 When using the backend-agnostic :class:`_functions.aggregate_strings` 

2133 string aggregation function, use the 

2134 :paramref:`_functions.aggregate_strings.order_by` parameter to indicate a 

2135 dialect-agnostic ORDER BY expression. 

2136 

2137 .. versionadded:: 2.1 Generalized the PostgreSQL-specific 

2138 :func:`_postgresql.aggregate_order_by` function to a method on 

2139 :class:`.Function` that is backend agnostic. 

2140 

2141 .. seealso:: 

2142 

2143 :class:`_functions.aggregate_strings` - backend-agnostic string 

2144 concatenation function which also supports ORDER BY 

2145 

2146 """ # noqa: E501 

2147 return AggregateOrderBy(element, *order_by)