Added new Core feature :func:`_sql.from_dml_column` that may be used in
expressions inside of :meth:`.UpdateBase.values` for INSERT or UPDATE; this
construct will copy whatever SQL expression is used for the given target
column in the statement to be used with additional columns. The construct
is mostly intended to be a helper with ORM :class:`.hybrid_property` within
DML hooks.
This is the Core side of the feature being added to the ORM
for #12496
Change-Id: Ic568638a8ce3607deea44af988b6451b30cde36c
manually update the files to remove literal string concat on the same line,
since black does not seem to be making progress in handling these
Change-Id: I3c651374c5f3db5b8bc0c700328d67ca03743b7b
Repaired a major shortcoming which was identified in the
:ref:`engine_insertmanyvalues` performance optimization feature first
introduced in the 2.0 series. This was a continuation of the change in
2.0.9 which disabled the SQL Server version of the feature due to a
reliance in the ORM on apparent row ordering that is not guaranteed to take
place. The fix applies new logic to all "insertmanyvalues" operations,
which takes effect when a new parameter
:paramref:`_dml.Insert.returning.sort_by_parameter_order` on the
:meth:`_dml.Insert.returning` or :meth:`_dml.UpdateBase.return_defaults`
methods, that through a combination of alternate SQL forms, direct
correspondence of client side parameters, and in some cases downgrading to
running row-at-a-time, will apply sorting to each batch of returned rows
using correspondence to primary key or other unique values in each row
which can be correlated to the input data.
Performance impact is expected to be minimal as nearly all common primary
key scenarios are suitable for parameter-ordered batching to be
achieved for all backends other than SQLite, while "row-at-a-time"
mode operates with a bare minimum of Python overhead compared to the very
heavyweight approaches used in the 1.x series. For SQLite, there is no
difference in performance when "row-at-a-time" mode is used.
It's anticipated that with an efficient "row-at-a-time" INSERT with
RETURNING batching capability, the "insertmanyvalues" feature can be later
be more easily generalized to third party backends that include RETURNING
support but not necessarily easy ways to guarantee a correspondence
with parameter order.
Fixes: #9618
References: #9603
Change-Id: I1d79353f5f19638f752936ba1c35e4dc235a8b7c
Fixed bug / regression where using :func:`.bindparam()` with the same name
as a column in the :meth:`.Update.values` method of :class:`.Update`, as
well as the :meth:`.Insert.values` method of :class:`.Insert` in 2.0 only,
would in some cases silently fail to honor the SQL expression in which the
parameter were presented, replacing the expression with a new parameter of
the same name and discarding any other elements of the SQL expression, such
as SQL functions, etc. The specific case would be statements that were
constructed against ORM entities rather than plain :class:`.Table`
instances, but would occur if the statement were invoked with a
:class:`.Session` or a :class:`.Connection`.
:class:`.Update` part of the issue was present in both 2.0 and 1.4 and is
backported to 1.4.
Fixes: #9075
Change-Id: Ie954bc1f492ec6a566163588182ef4910c7ee452
Changed how the positional compilation is performed. It's rendered by the compiler
the same as the pyformat compilation. The string is then processed to replace
the placeholders with the correct ones, and to obtain the correct order of the
parameters.
This vastly simplifies the computation of the order of the parameters, that in
case of nested CTE is very hard to compute correctly.
Reworked how numeric paramstyle behavers:
- added support for repeated parameter, without duplicating them like in normal
positional dialects
- implement insertmany support. This requires that the dialect supports out of
order placehoders, since all parameters that are not part of the VALUES clauses
are placed at the beginning of the parameter tuple
- support for different identifiers for a numeric parameter. It's for example
possible to use postgresql style placeholder $1, $2, etc
Added two new dialect based on sqlite to test "numeric" fully using
both :1 style and $1 style. Includes a workaround for SQLite's
not-really-correct numeric implementation.
Changed parmstyle of asyncpg dialect to use numeric, rendering with its native
$ identifiers
Fixes: #8926Fixes: #8849
Change-Id: I7c640467d49adfe6d795cc84296fc7403dcad4d6
command run is "pyupgrade --py37-plus --keep-runtime-typing --keep-percent-format <files...>"
pyupgrade will change assert_ to assertTrue. That was reverted since assertTrue does not
exists in sqlalchemy fixtures
Change-Id: Ie1ed2675c7b11d893d78e028aad0d1576baebb55
The :class:`.Sequence` construct restores itself to the DDL behavior it
had prior to the 1.4 series, where creating a :class:`.Sequence` with
no additional arguments will emit a simple ``CREATE SEQUENCE`` instruction
**without** any additional parameters for "start value". For most backends,
this is how things worked previously in any case; **however**, for
MS SQL Server, the default value on this database is
``-2**63``; to prevent this generally impractical default
from taking effect on SQL Server, the :paramref:`.Sequence.start` parameter
should be provided. As usage of :class:`.Sequence` is unusual
for SQL Server which for many years has standardized on ``IDENTITY``,
it is hoped that this change has minimal impact.
Fixes: #7211
Change-Id: I1207ea10c8cb1528a1519a0fb3581d9621c27b31
the feature is enabled for all built in backends
when RETURNING is used,
except for Oracle that doesn't need it, and on
psycopg2 and mssql+pyodbc it is used for all INSERT statements,
not just those that use RETURNING.
third party dialects would need to opt in to the new feature
by setting use_insertmanyvalues to True.
Also adds dialect-level guards against using returning
with executemany where we dont have an implementation to
suit it. execute single w/ returning still defers to the
server without us checking.
Fixes: #6047Fixes: #7907
Change-Id: I3936d3c00003f02e322f2e43fb949d0e6e568304
Fixed bug in the behavior of the :paramref:`_orm.Mapper.eager_defaults`
parameter such that client-side SQL default or onupdate expressions in the
table definition alone will trigger a fetch operation using RETURNING or
SELECT when the ORM emits an INSERT or UPDATE for the row. Previously, only
server side defaults established as part of table DDL and/or server-side
onupdate expressions would trigger this fetch, even though client-side SQL
expressions would be included when the fetch was rendered.
Fixes: #7438
Change-Id: Iba719298ba4a26d185edec97ba77d2d54585e5a4
just in my own testing, if I say insert().return_defaults()
and stringify, I should see it, so make sure all the dialects
default to "insert_returning" etc. , with downgrade on
server version check.
Change-Id: Id64e78fcb03c48b5dcb0feb21cb9cc495edd15e9
As almost every dialect supports RETURNING now, RETURNING
is also made more of a default assumption.
* the default compiler generates a RETURNING clause now
when specified; CompileError is no longer raised.
* The dialect-level implicit_returning parameter now has
no effect. It's not fully clear if there are real world
cases relying on the dialect-level parameter, so we will see
once 2.0 is released. ORM-level RETURNING can be disabled
at the table level, and perhaps "implicit returning" should
become an ORM-level option at some point as that's where
it applies.
* Altered ORM update() / delete() to respect table-level
implicit returning for fetch.
* Since MariaDB doesnt support UPDATE returning, "full_returning"
is now split into insert_returning, update_returning, delete_returning
* Crazy new thing. Dialects that have *both* cursor.lastrowid
*and* returning. so now we can pick between them for SQLite
and mariadb. so, we are trying to keep it on .lastrowid for
simple inserts with an autoincrement column, this helps with
some edge case test scenarios and i bet .lastrowid is faster
anyway. any return_defaults() / multiparams etc then we
use returning
* SQLite decided they dont want to return rows that match in
ON CONFLICT. this is flat out wrong, but for now we need to
work with it.
Fixes: #6195Fixes: #7011Closes: #7047
Pull-request: https://github.com/sqlalchemy/sqlalchemy/pull/7047
Pull-request-sha: d25d5ea3ab
Co-authored-by: Mike Bayer <mike_mp@zzzcomputing.com>
Change-Id: I9908ce0ff7bdc50bd5b27722081767c31c19a950
An informative error is raised for the use case where
:meth:`.Insert.from_select` is being passed a "compound select" object such
as a UNION, yet the INSERT statement needs to append additional columns to
support Python-side or explicit SQL defaults from the table metadata. In
this case a subquery of the compound object should be passed.
Fixes: #8073
Change-Id: Ic4a5dbf84ec49d2451901be05cb9cf6ae93f02b7
Altered the compilation mechanics of the :class:`.Insert` construct such
that the "autoincrement primary key" column value will be fetched via
``cursor.lastrowid`` or RETURNING even if present in the parameter set or
within the :meth:`.Insert.values` method as a plain bound value, for
single-row INSERT statements on specific backends that are known to
generate autoincrementing values even when explicit NULL is passed. This
restores a behavior that was in the 1.3 series for both the use case of
separate parameter set as well as :meth:`.Insert.values`. In 1.4, the
parameter set behavior unintentionally changed to no longer do this, but
the :meth:`.Insert.values` method would still fetch autoincrement values up
until 1.4.21 where 🎫`6770` changed the behavior yet again again
unintentionally as this use case was never covered.
The behavior is now defined as "working" to suit the case where databases
such as SQLite, MySQL and MariaDB will ignore an explicit NULL primary key
value and nonetheless invoke an autoincrement generator.
Fixes: #7998
Change-Id: I5d4105a14217945f87fbe9a6f2a3c87f6ef20529
Full "RETURNING" support is implemented for the cx_Oracle dialect, meaning
multiple RETURNING rows are now recived for DML statements that produce
more than one row for RETURNING.
cx_Oracle 7 is now the minimum version for cx_Oracle.
Getting Oracle to do multirow returning took about 5 minutes. however,
getting Oracle's RETURNING system to integrate with ORM-enabled
insert, update, delete, is a big deal because that architecture wasn't
really working very robustly, including some recent changes in 1.4
for FromStatement were done in a hurry, so this patch also cleans up
the FromStatement situation and begins to establish it more concretely
as the base for all ReturnsRows / TextClause ORM scenarios.
Fixes: #6245
Change-Id: I2b4e6007affa51ce311d2d5baa3917f356ab961f
Fixed issue where using ORM column expressions as keys in the list of
dictionaries passed to :meth:`_sql.Insert.values` for "multi-valued insert"
would not be processed correctly into the correct column expressions.
Fixes: #7060
Change-Id: I1c4c286c33ea6eeaafba617996828f5c88ff0a1c
Fixed issue where usage of an explicit :class:`.Sequence` would produce
inconsistent "inline" behavior for an :class:`.Insert` construct that
includes multiple values phrases; the first seq would be inline but
subsequent ones would be "pre-execute", leading to inconsistent sequence
ordering. The sequence expressions are now fully inline.
Fixes: #6361
Change-Id: Ie16794ec0e19979a7e6c8d1bef5716a9fc199889
Fixed regression where the introduction of the INSERT syntax "INSERT...
VALUES (DEFAULT)" was not supported on some backends that do however
support "INSERT..DEFAULT VALUES", including SQLite. The two syntaxes are
now each individually supported or non-supported for each dialect, for
example MySQL supports "VALUES (DEFAULT)" but not "DEFAULT VALUES".
Support for Oracle is still not enabled as there are unresolved issues
in using RETURNING at the same time.
Fixes: #6254
Change-Id: I47959bc826e3d9d2396ccfa290eb084841b02e77
Adjusted the "literal_binds" feature of :class:`_sql.Compiler` to render
NULL for a bound parameter that has ``None`` as the value, either
explicitly passed or omitted. The previous error message "bind parameter
without a renderable value" is removed, and a missing or ``None`` value
will now render NULL in all cases. Previously, rendering of NULL was
starting to happen for DML statements due to internal refactorings, but was
not explicitly part of test coverage, which it now is.
While no error is raised, when the context is within that of a column
comparison, and the operator is not "IS"/"IS NOT", a warning is emitted
that this is not generally useful from a SQL perspective.
Fixes: #5888
Change-Id: Id5939d8dbfb1156a9f8a7f7e76cf18327155331a
As the test suite has widespread use of many patterns
that are deprecated, enable SQLALCHEMY_WARN_20 globally
for the test suite but then break the warnings filter
out into a whole list of all the individual warnings
we are looking for. this way individual changesets
can target a specific class of warning, as many of these
warnings will indivdidually affect dozens of files
and potentially hundreds of lines of code.
Many warnings are also resolved here as this
patch started out that way. From this point
forward there should be changesets that target a
subset of the warnings at a time.
For expediency, updates some migration 2.0 docs
for ORM as well.
Change-Id: I98b8defdf7c37b818b3824d02f7668e3f5f31c94
This change includes mainly that the bracketed use within
select() is moved to positional, and keyword arguments are
removed from calls to the select() function. it does not
yet fully address other issues such as keyword arguments passed
to the table.select().
Additionally, allows False / None to both be considered
as "disable" for all of select.correlate(), select.correlate_except(),
query.correlate(), which establishes consistency with
passing of ``False`` for the legact select(correlate=False)
argument.
Change-Id: Ie6c6e6abfbd3d75d4c8de504c0cf0159e6999108
Build on #5401 to allow the ORM to take advanage
of executemany INSERT + RETURNING.
Implemented the feature
updated tests
to support INSERT DEFAULT VALUES, needed to come up with
a new syntax for compiler INSERT INTO table (anycol) VALUES (DEFAULT)
which can then be iterated out for executemany.
Added graceful degrade to plain executemany for PostgreSQL <= 8.2
Renamed EXECUTEMANY_DEFAULT to EXECUTEMANY_PLAIN
Fix issue where unicode identifiers or parameter names wouldn't
work with execute_values() under Py2K, because we have to
encode the statement and therefore have to encode the
insert_single_values_expr too.
Correct issue from #5401 to support executemany + return_defaults
for a PK that is explicitly pre-generated, meaning we aren't actually
getting RETURNING but need to return it from compiled_parameters.
Fixes: #5263
Change-Id: Id68e5c158c4f9ebc33b61c06a448907921c2a657
The psycopg2 dialect now defaults to using the very performant
``execute_values()`` psycopg2 extension for compiled INSERT statements,
and also impements RETURNING support when this extension is used. This
allows INSERT statements that even include an autoincremented SERIAL
or IDENTITY value to run very fast while still being able to return the
newly generated primary key values. The ORM will then integrate this
new feature in a separate change.
Implements RETURNING for insert with executemany
Adds support to return_defaults() mode and inserted_primary_key
to support mutiple INSERTed rows, via return_defauls_rows
and inserted_primary_key_rows accessors.
within default execution context, new cached compiler
getters are used to fetch primary keys from rows
inserted_primary_key now returns a plain tuple. this
is not yet a row-like object however this can be
added.
Adds distinct "values_only" and "batch" modes, as
"values" has a lot of benefits but "batch" breaks
cursor.rowcount
psycopg2 minimum version 2.7 so we can remove the
large number of checks for very old versions of
psycopg2
simplify tests to no longer distinguish between
native and non-native json
Fixes: #5401
Change-Id: Ic08fd3423d4c5d16ca50994460c0c234868bd61c
Targeting select / insert / update / delete, the goal
is to minimize overhead of construction and generative methods
so that only the raw arguments passed are handled. An interim
stage that converts the raw state into more compiler-ready state
is added, which is analogous to the ORM QueryContext which will
also be rolled in to be a similar concept, as is currently
being prototyped in I19e05b3424b07114cce6c439b05198ac47f7ac10.
the ORM update/delete BulkUD concept is also going to be rolled
onto this idea. So while the compiler-ready state object,
here called DMLState, looks a little thin, it's the
base of a bigger pattern that will allow for ORM functionality
to embed itself directly into the compiler, execution
context, and result set objects.
This change targets the DML objects, primarily focused on the
values() method which is the most complex process. The
work done by values() is minimized as much as possible
while still being able to create a cache key. Additional
computation is then offloaded to a new object ValuesState
that is handled by the compiler.
Architecturally, a big change here is that insert.values()
and update.values() will generate BindParameter objects for
the values now, which are then carefully received by crud.py
so that they generate the expected names. This is so that
the values() portion of these constructs is cacheable.
for the "multi-values" version of Insert, this is all skipped
and the plan right now is that a multi-values insert is
not worth caching (can always be revisited).
Using the
coercions system in values() also gets us nicer validation
for free, we can remove the NotAClauseElement thing from
schema, and we also now require scalar_subquery() is called
for an insert/update that uses a SELECT as a column value,
1.x deprecation path is added.
The traversal system is then applied to the DML objects
including tests so that they have traversal, cloning, and
cache key support. cloning is not a use case for DML however
having it present allows better validation of the structure
within the tests.
Special per-dialect DML is explicitly not cacheable at the moment,
more as a proof of concept that third party DML constructs can
exist as gracefully not-cacheable rather than producing an
incomplete cache key.
A few selected performance improvements have been added as well,
simplifying the immutabledict.union() method and adding
a new SQLCompiler function that can generate delimeter-separated
clauses like WHERE and ORDER BY without having to build
a ClauseList object at all. The use of ClauseList will
be removed from Select in an upcoming commit. Overall,
ClaustList is unnecessary for internal use and only adds
overhead to statement construction and will likely be removed
as much as possible except for explcit use of conjunctions like
and_() and or_().
Change-Id: I408e0b8be91fddd77cf279da97f55020871f75a9
In 9fca5d827d we attempted to deprecate the "inline=True" flag
and add a generative inline() method, however failed to include
any tests and the method was implemented incorrectly such that
it would get overwritten with the boolean flag immediately.
Rename the internal "inline" flag to "_inline" and add test
support both for the method as well as deprecated support
for the flag, including a fixture addition to assert the expected
value of the flag as it generally does not affect the
actual compiled SQL string.
Change-Id: I0450049f17f1f0d91e22d27f1a973a2b6c0e59f7
A SQL expression can now be assigned to a primary key attribute for an ORM
flush in the same manner as ordinary attributes as described in
:ref:`flush_embedded_sql_expressions` where the expression will be evaulated
and then returned to the ORM using RETURNING, or in the case of pysqlite,
works using the cursor.lastrowid attribute.Requires either a database that
supports RETURNING (e.g. Postgresql, Oracle, SQL Server) or pysqlite.
Fixes: #3133Fixes: #4494
Change-Id: I83da8357354de002cb04fa4a553f2a2f90c5157d
Applied on top of a pure run of black -l 79 in
I7eda77fed3d8e73df84b3651fd6cfcfe858d4dc9, this set of changes
resolves all remaining flake8 conditions for those codes
we have enabled in setup.cfg.
Included are resolutions for all remaining flake8 issues
including shadowed builtins, long lines, import order, unused
imports, duplicate imports, and docstring issues.
Change-Id: I4f72d3ba1380dd601610ff80b8fb06a2aff8b0fe
This is a straight reformat run using black as is, with no edits
applied at all.
The black run will format code consistently, however in
some cases that are prevalent in SQLAlchemy code it produces
too-long lines. The too-long lines will be resolved in the
following commit that will resolve all remaining flake8 issues
including shadowed builtins, long lines, import order, unused
imports, duplicate imports, and docstring issues.
Change-Id: I7eda77fed3d8e73df84b3651fd6cfcfe858d4dc9
test_compiler is mostly related to SELECT statements as well
as smaller SQL elements. While it still has some DDL related
tests, move out all the remaining insert/update tests into
the already present test_insert.py, test_update.py
Fixes: #2630
Change-Id: I4167618543fd1235d12d1717c8c629d2374b325a
Fixed INSERT FROM SELECT with CTEs for the Oracle and MySQL dialects, where
the CTE was being placed above the entire statement as is typical with
other databases, however Oracle and MariaDB 10.2 wants the CTE underneath
the "INSERT" segment. Note that the Oracle and MySQL dialects don't yet
work when a CTE is applied to a subquery inside of an UPDATE or DELETE
statement, as the CTE is still applied to the top rather than inside the
subquery.
Also adds test suite support CTEs against backends.
Change-Id: I8ac337104d5c546dd4f0cd305632ffb56ac8bf90
Fixes: #4275Fixes: #4230
Fixed bug in :meth:`.Insert.values` where using the "multi-values"
format in combination with :class:`.Column` objects as keys rather
than strings would fail. Pull request courtesy Aubrey Stark-Toller.
Change-Id: I9d3b40b5950df8f5bfdc8b1d22f9c3afb277f17f
Pull-request: https://github.com/zzzeek/sqlalchemy/pull/412Fixes: #4162
Fixed bug where literal_binds compiler flag was not honored by the
:class:`.Insert` construct for the "multiple values" feature; the
subsequent values are now rendered as literals.
Change-Id: I81ac358fd59995885d482e7571620090210865d2
Fixes: #3880
Fixed bug where an INSERT from SELECT where the source table contains
an autoincrementing Sequence would fail to compile correctly.
Change-Id: I41eb9f65789a4007712ae61ed5fa23a9839a5128
Fixes: #3877
PrimaryKeyConstraint is present on Table however
on table() and others it's a ColumnSet. The warning
here only needs len() and PrimaryKeyConstraint supports that
directly in the same way as ColumnSet.
Change-Id: I19c11a39110bfef48cdea49a471e7ab80b537538
Fixes: #3842
Changed the naming convention used when generating bound parameters
for a multi-VALUES insert statement, so that the numbered parameter
names don't conflict with the anonymized parameters of a WHERE clause,
as is now common in a PostgreSQL ON CONFLICT construct.
Change-Id: I3188d100fe4d322a47d344d6a63d3e40b915f228
Fixes: #3828
INSERT, UPDATE, and DELETE statements to both specify their own
WITH clause, as well as for these statements themselves to be
CTE expressions when they include a RETURNING clause.
fixes#2551
the :class:`.Select` construct would have its ``._raw_columns``
collection mutated in-place when compiling the :class:`.Insert`
construct, when the target :class:`.Table` has Python-side defaults.
The :class:`.Select` construct would compile standalone with the
erroneous column present subsequent to compilation of the
:class:`.Insert`, and the the :class:`.Insert` statement itself would
fail on a second compile attempt due to duplicate bound parameters.
fixes#3603
"auto increment" column has been changed, such that autoincrement
is no longer implicitly enabled for a :class:`.Table` that has a
composite primary key. In order to accommodate being able to enable
autoincrement for a composite PK member column while at the same time
maintaining SQLAlchemy's long standing behavior of enabling
implicit autoincrement for a single integer primary key, a third
state has been added to the :paramref:`.Column.autoincrement` parameter
``"auto"``, which is now the default. fixes#3216
- The MySQL dialect no longer generates an extra "KEY" directive when
generating CREATE TABLE DDL for a table using InnoDB with a
composite primary key with AUTO_INCREMENT on a column that isn't the
first column; to overcome InnoDB's limitation here, the PRIMARY KEY
constraint is now generated with the AUTO_INCREMENT column placed
first in the list of columns.
insert statement, 🎫`3288`, where the column type for the
default-holding column would not be propagated to the compiled
statement in the case where the default was being used,
leading to bind-level type handlers not being invoked.
fixes#3520
inside of :meth:`.Insert.from_select`. This behavior worked
accidentally up until 0.9.9, when it no longer worked due to
unrelated changes as part of 🎫`3248`. Note that this
is the rendering of the WITH clause after the INSERT, before the
SELECT; the full functionality of CTEs rendered at the top
level of INSERT, UPDATE, DELETE is a new feature targeted for a
later release.
fixes#3418
repaired to work more usefully with tables that have Python-
side default values and/or functions, as well as server-side
defaults. The feature will now work with a dialect that uses
"positional" parameters; a Python callable will also be
invoked individually for each row just as is the case with an
"executemany" style invocation; a server- side default column
will no longer implicitly receive the value explicitly
specified for the first row, instead refusing to invoke
without an explicit value. fixes#3288
defaults if otherwise unspecified; the limitation where non-
server column defaults aren't included in an INSERT FROM
SELECT is now lifted and these expressions are rendered as
constants into the SELECT statement.
constructs are now importable from the "from sqlalchemy" namespace,
just like every other Core construct.
- The implicit conversion of strings to :func:`.text` constructs
when passed to most builder methods of :func:`.select` as
well as :class:`.Query` now emits a warning with just the
plain string sent. The textual conversion still proceeds normally,
however. The only method that accepts a string without a warning
are the "label reference" methods like order_by(), group_by();
these functions will now at compile time attempt to resolve a single
string argument to a column or label expression present in the
selectable; if none is located, the expression still renders, but
you get the warning again. The rationale here is that the implicit
conversion from string to text is more unexpected than not these days,
and it is better that the user send more direction to the Core / ORM
when passing a raw string as to what direction should be taken.
Core/ORM tutorials have been updated to go more in depth as to how text
is handled.
fixes#2992
Mike Bayer