Mike Bayer [Sat, 19 Sep 2015 22:06:23 +0000 (18:06 -0400)]
- The ``legacy_schema_aliasing`` flag, introduced in version 1.0.5
as part of :ticket:`3424` to allow disabling of the MSSQL dialect's
attempts to create aliases for schema-qualified tables, now defaults
to False; the old behavior is now disabled unless explicitly turned on.
fixes #3434
Mike Bayer [Sat, 19 Sep 2015 17:12:08 +0000 (13:12 -0400)]
- Added a new type-level modifier :meth:`.TypeEngine.evaluates_none`
which indicates to the ORM that a positive set of None should be
persisted as the value NULL, instead of omitting the column from
the INSERT statement. This feature is used both as part of the
implementation for :ticket:`3514` as well as a standalone feature
available on any type. fixes #3250
- add new documentation section illustrating the "how to force null"
use case of #3250
- alter our change from #3514 so that the class-level flag is now
called "should_evaluate_none"; so that "evaluates_none" is now
a generative method.
Mike Bayer [Sat, 19 Sep 2015 15:58:50 +0000 (11:58 -0400)]
- The use of a :class:`.postgresql.ARRAY` object that refers
to a :class:`.types.Enum` or :class:`.postgresql.ENUM` subtype
will now emit the expected "CREATE TYPE" and "DROP TYPE" DDL when
the type is used within a "CREATE TABLE" or "DROP TABLE".
fixes #2729
Mike Bayer [Wed, 16 Sep 2015 22:46:53 +0000 (18:46 -0400)]
- The :func:`.type_coerce` construct is now a fully fledged Core
expression element which is late-evaluated at compile time. Previously,
the function was only a conversion function which would handle different
expression inputs by returning either a :class:`.Label` of a column-oriented
expression or a copy of a given :class:`.BindParameter` object,
which in particular prevented the operation from being logically
maintained when an ORM-level expression transformation would convert
a column to a bound parameter (e.g. for lazy loading).
fixes #3531
Mike Bayer [Thu, 10 Sep 2015 14:00:46 +0000 (10:00 -0400)]
- add a new FAQ recipe for "walk all objects", replacing the need
to use mapper.cascade_iterator() for this purpose as it was not really
designed for that use case. Add docs to cascade_iterator() pointing
to the recipe. fixes #3498
Mike Bayer [Tue, 8 Sep 2015 18:12:32 +0000 (14:12 -0400)]
- Internal calls to "bookkeeping" functions within
:meth:`.Session.bulk_save_objects` and related bulk methods have
been scaled back to the extent that this functionality is not
currently used, e.g. checks for column default values to be
fetched after an INSERT or UPDATE statement.
fixes #3526
Mike Bayer [Tue, 8 Sep 2015 17:00:26 +0000 (13:00 -0400)]
- Fixed bug in :meth:`.Session.bulk_save_objects` where a mapped
column that had some kind of "fetch on update" value and was not
locally present in the given object would cause an AttributeError
within the operation.
fixes #3525
Mike Bayer [Fri, 28 Aug 2015 21:43:46 +0000 (17:43 -0400)]
- The :class:`.SessionEvents` suite now includes events to allow
unambiguous tracking of all object lifecycle state transitions
in terms of the :class:`.Session` itself, e.g. pending,
transient, persistent, detached. The state of the object
within each event is also defined.
fixes #2677
- Added a new session lifecycle state :term:`deleted`. This new state
represents an object that has been deleted from the :term:`persistent`
state and will move to the :term:`detached` state once the transaction
is committed. This resolves the long-standing issue that objects
which were deleted existed in a gray area between persistent and
detached. The :attr:`.InstanceState.persistent` accessor will
**no longer** report on a deleted object as persistent; the
:attr:`.InstanceState.deleted` accessor will instead be True for
these objects, until they become detached.
- The :paramref:`.Session.weak_identity_map` parameter is deprecated.
See the new recipe at :ref:`session_referencing_behavior` for
an event-based approach to maintaining strong identity map behavior.
references #3517
Mike Bayer [Mon, 31 Aug 2015 15:30:03 +0000 (11:30 -0400)]
- Fixed regression in 1.0-released default-processor for multi-VALUES
insert statement, :ticket:`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
Mike Bayer [Fri, 28 Aug 2015 03:38:14 +0000 (23:38 -0400)]
- use consistent and descriptive language in all cases
where we refer to the "weak_identity_map" option, and add additional
exposition in the session documentation which refers to it.
fixes #3517
Mike Bayer [Thu, 27 Aug 2015 22:04:25 +0000 (18:04 -0400)]
- The :class:`.TypeDecorator` type extender will now work in conjunction
with a :class:`.SchemaType` implementation, typically :class:`.Enum`
or :class:`.Boolean` with regards to ensuring that the per-table
events are propagated from the implementation type to the outer type.
These events are used
to ensure that the constraints or Postgresql types (e.g. ENUM)
are correctly created (and possibly dropped) along with the parent
table.
fixes #2919
Mike Bayer [Thu, 27 Aug 2015 19:56:00 +0000 (15:56 -0400)]
- reword and update documentation on passive_updates=False; as virtually
all DBs now support ON UPDATE CASCADE other than Oracle, there's no need
to try to reimplement more functionality here and users should be encouraged
to forego natural PKs unless their target platform supports them.
references #2666
Mike Bayer [Wed, 26 Aug 2015 20:58:13 +0000 (16:58 -0400)]
- Added support for "set-aggregate" functions of the form
``<function> WITHIN GROUP (ORDER BY <criteria>)``, using the
method :class:`.FunctionElement.within_group`. A series of common
set-aggregate functions with return types derived from the set have
been added. This includes functions like :class:`.percentile_cont`,
:class:`.dense_rank` and others.
fixes #1370
- make sure we use func.name for all _literal_as_binds in functions.py
so we get consistent naming behavior for parameters.
Mike Bayer [Wed, 26 Aug 2015 19:15:45 +0000 (15:15 -0400)]
- Added support for the SQL-standard function :class:`.array_agg`,
which automatically returns an :class:`.Array` of the correct type
and supports index / slice operations. As arrays are only
supported on Postgresql at the moment, only actually works on
Postgresql. fixes #3132
Mike Bayer [Mon, 24 Aug 2015 21:57:36 +0000 (17:57 -0400)]
- build out a new base type for Array, as well as new any/all operators
- any/all work for Array as well as subqueries, accepted by MySQL
- Postgresql ARRAY now subclasses Array
- fixes #3516
Mike Bayer [Mon, 24 Aug 2015 14:54:09 +0000 (10:54 -0400)]
- Fixed two issues in the "history_meta" example where history tracking
could encounter empty history, and where a column keyed to an alternate
attribute name would fail to track properly. Fixes courtesy
Alex Fraser.
Mike Bayer [Sat, 22 Aug 2015 16:47:13 +0000 (12:47 -0400)]
- Added new checks for the common error case of passing mapped classes
or mapped instances into contexts where they are interpreted as
SQL bound parameters; a new exception is raised for this.
fixes #3321
Mike Bayer [Fri, 21 Aug 2015 20:36:15 +0000 (16:36 -0400)]
- rework the "controlling DDL sequences" documentation to
refer mostly to the DDL object; this system is primarily useful
in that case, and not for built-in objects. Reference that
the built-in case is not really viable right now. References #3442.
Mike Bayer [Tue, 18 Aug 2015 17:02:58 +0000 (13:02 -0400)]
- fix the postgresql_jsonb requirement to include the 9.4 requirement
- new test for json col['x']['y']['z'] seems to fail pre PG 9.4,
fails on comparisons for non-compatible data instead of not matching
- no need to call SpecPredicate(db) directly in exclusion functions,
by using Predicate.as_predicate() the spec strings can have version
comparisons
Mike Bayer [Tue, 18 Aug 2015 16:40:18 +0000 (12:40 -0400)]
- as the Concatenable mixin was changed to support calling down to
"super" instead of hardcoding to "self.type" for the default return
value, the base Comparator was returning other_comparator.type. It's
not clear what the rationale for this was, though in theory the
base Comparator should possibly even throw an exception if the two
types aren't the same (or of the same affinity?) .
- mysql.SET was broken on this because the bitwise version adds "0"
to the value to force an integer within column_expression, we are doing type_coerces here
now in any case so that there is no type ambiguity for this
operation
Mike Bayer [Mon, 17 Aug 2015 21:04:33 +0000 (17:04 -0400)]
- merge of ticket_3514 None-handling branch
- Fixes to the ORM and to the postgresql JSON type regarding the
``None`` constant in conjunction with the Postgresql :class:`.JSON` type. When
the :paramref:`.JSON.none_as_null` flag is left at its default
value of ``False``, the ORM will now correctly insert the Json
"'null'" string into the column whenever the value on the ORM
object is set to the value ``None`` or when the value ``None``
is used with :meth:`.Session.bulk_insert_mappings`,
**including** if the column has a default or server default on it. This
makes use of a new type-level flag "evaluates_none" which is implemented
by the JSON type based on the none_as_null flag. fixes #3514
- Added a new constant :attr:`.postgresql.JSON.NULL`, indicating
that the JSON NULL value should be used for a value
regardless of other settings. part of fixes #3514
Mike Bayer [Mon, 17 Aug 2015 20:43:54 +0000 (16:43 -0400)]
- merge of ticket_3499 indexed access branch
- The "hashable" flag on special datatypes such as :class:`.postgresql.ARRAY`,
:class:`.postgresql.JSON` and :class:`.postgresql.HSTORE` is now
set to False, which allows these types to be fetchable in ORM
queries that include entities within the row. fixes #3499
- The Postgresql :class:`.postgresql.ARRAY` type now supports multidimensional
indexed access, e.g. expressions such as ``somecol[5][6]`` without
any need for explicit casts or type coercions, provided
that the :paramref:`.postgresql.ARRAY.dimensions` parameter is set to the
desired number of dimensions. fixes #3487
- The return type for the :class:`.postgresql.JSON` and :class:`.postgresql.JSONB`
when using indexed access has been fixed to work like Postgresql itself,
and returns an expression that itself is of type :class:`.postgresql.JSON`
or :class:`.postgresql.JSONB`. Previously, the accessor would return
:class:`.NullType` which disallowed subsequent JSON-like operators to be
used. part of fixes #3503
- The :class:`.postgresql.JSON`, :class:`.postgresql.JSONB` and
:class:`.postgresql.HSTORE` datatypes now allow full control over the
return type from an indexed textual access operation, either ``column[someindex].astext``
for a JSON type or ``column[someindex]`` for an HSTORE type,
via the :paramref:`.postgresql.JSON.astext_type` and
:paramref:`.postgresql.HSTORE.text_type` parameters. also part of fixes #3503
- The :attr:`.postgresql.JSON.Comparator.astext` modifier no longer
calls upon :meth:`.ColumnElement.cast` implicitly, as PG's JSON/JSONB
types allow cross-casting between each other as well. Code that
makes use of :meth:`.ColumnElement.cast` on JSON indexed access,
e.g. ``col[someindex].cast(Integer)``, will need to be changed
to call :attr:`.postgresql.JSON.Comparator.astext` explicitly. This is
part of the refactor in references #3503 for consistency in operator
use.
Mike Bayer [Fri, 14 Aug 2015 19:34:01 +0000 (15:34 -0400)]
- add tests for InstanceEvents.init, InstanceEvents.init_failure
- ensure that kwargs can be modified in-place within InstanceEvents.init
and that these take effect for the __init__ method.
- improve documentation for these and related events, including
that kwargs can be modified in-place.
Donald Stufft [Thu, 13 Aug 2015 18:57:53 +0000 (14:57 -0400)]
Tell Wheel to generate platform and interpreter specific wheels
By telling wheel that we have extension modules, even though we
have none, wheel will create a Wheel which is platform and
interpreter specific. This will ensure that the pure Python wheels
on PyPy do not trigger installs on CPython without the C speedups.
Mike Bayer [Wed, 12 Aug 2015 18:26:11 +0000 (14:26 -0400)]
- The behavior of the :func:`.union` construct and related constructs
such as :meth:`.Query.union` now handle the case where the embedded
SELECT statements need to be parenthesized due to the fact that they
include LIMIT, OFFSET and/or ORDER BY. These queries **do not work
on SQLite**, and will fail on that backend as they did before, but
should now work on all other backends.
fixes #2528
Mike Bayer [Tue, 11 Aug 2015 17:05:17 +0000 (13:05 -0400)]
- Fixed 1.0 regression where the "noload" loader strategy would fail
to function for a many-to-one relationship. The loader used an
API to place "None" into the dictionary which no longer actually
writes a value; this is a side effect of :ticket:`3061`.
- remove InstanceState._initialize() totally, it's used nowhere
else and no longer does what it says it does
- fill in fowards-port version ids throughout the changes for 1.0.9
Mike Bayer [Mon, 10 Aug 2015 14:07:17 +0000 (10:07 -0400)]
- Fixed two issues regarding Sybase reflection, allowing tables
without primary keys to be reflected as well as ensured that
a SQL statement involved in foreign key detection is pre-fetched up
front to avoid driver issues upon nested queries. Fixes here
courtesy Eugene Zapolsky; note that we cannot currently test
Sybase to locally verify these changes.
fixes #3508 fixes #3509
Mike Bayer [Thu, 30 Jul 2015 14:34:36 +0000 (10:34 -0400)]
- Fixed issue where the SQL Server dialect would reflect a string-
or other variable-length column type with unbounded length
by assigning the token ``"max"`` to the
length attribute of the string. While using the ``"max"`` token
explicitly is supported by the SQL Server dialect, it isn't part
of the normal contract of the base string types, and instead the
length should just be left as None. The dialect now assigns the
length to None on reflection of the type so that the type behaves
normally in other contexts.
fixes #3504
Mike Bayer [Wed, 29 Jul 2015 21:40:47 +0000 (17:40 -0400)]
- Fixed support for cx_Oracle version 5.2, which was tripping
up SQLAlchemy's version detection under Python 3 and inadvertently
not using the correct unicode mode for Python 3. This would cause
issues such as bound variables mis-interpreted as NULL and rows
silently not being returned.
fixes #3491
Mike Bayer [Fri, 24 Jul 2015 22:52:25 +0000 (18:52 -0400)]
- remove "distutils-only" mode, use setuptools only
- add first-class "pytest" runner to setup.py, fixes #3489
- remove references to setuptools Feature and mark that "--without-cextensions"
is removed, fixes #3500
Mike Bayer [Fri, 24 Jul 2015 16:24:18 +0000 (12:24 -0400)]
- An adjustment to the new Postgresql feature of reflecting storage
options and USING of :ticket:`3455` released in 1.0.6,
to disable the feature for Postgresql versions < 8.2 where the
``reloptions`` column is not provided; this allows Amazon Redshift
to again work as it is based on an 8.0.x version of Postgresql.
Fix courtesy Pete Hollobon.
references #3455
Mike Bayer [Wed, 22 Jul 2015 21:59:34 +0000 (17:59 -0400)]
- Fixed critical issue whereby the pool "checkout" event handler
may be called against a stale connection without the "connect"
event handler having been called, in the case where the pool
attempted to reconnect after being invalidated and failed; the stale
connection would remain present and would be used on a subsequent
attempt. This issue has a greater impact in the 1.0 series subsequent
to 1.0.2, as it also delivers a blanked-out ``.info`` dictionary to
the event handler; prior to 1.0.2 the ``.info`` dictionary is still
the previous one.
fixes #3497
Mike Bayer [Tue, 21 Jul 2015 16:33:35 +0000 (12:33 -0400)]
- Fixed bug in SQLite dialect where reflection of UNIQUE constraints
that included non-alphabetic characters in the names, like dots or
spaces, would not be reflected with their name.
fixes #3495
Mike Bayer [Tue, 21 Jul 2015 00:35:04 +0000 (20:35 -0400)]
- Fixed an issue where a particular base class within utils
didn't implement ``__slots__``, and therefore meant all subclasses
of that class didn't either, negating the rationale for ``__slots__``
to be in use. Didn't cause any issue except on IronPython
which apparently does not implement ``__slots__`` behavior compatibly
with cPython.
Fixes #3494
Mike Bayer [Sun, 19 Jul 2015 20:32:31 +0000 (16:32 -0400)]
- Fixed regression where new methods on :class:`.ResultProxy` used
by the ORM :class:`.Query` object (part of the performance
enhancements of :ticket:`3175`) would not raise the "this result
does not return rows" exception in the case where the driver
(typically MySQL) fails to generate cursor.description correctly;
an AttributeError against NoneType would be raised instead.
fixes #3481
Mike Bayer [Sun, 19 Jul 2015 16:20:00 +0000 (12:20 -0400)]
- Fixed regression where :meth:`.ResultProxy.keys` would return
un-adjusted internal symbol names for "anonymous" labels, which
are the "foo_1" types of labels we see generated for SQL functions
without labels and similar. This was a side effect of the
performance enhancements implemented as part of references #918.
fixes #3483
Mike Bayer [Sun, 19 Jul 2015 14:27:33 +0000 (10:27 -0400)]
- Fixed bug where coersion of literal ``True`` or ``False`` constant
in conjunction with :func:`.and_` or :func:`.or_` would fail
with an AttributeError.
fixes #3490
Mike Bayer [Sun, 19 Jul 2015 14:00:13 +0000 (10:00 -0400)]
- Fixed potential issue where a custom subclass
of :class:`.FunctionElement` or other column element that incorrectly
states 'None' or any other invalid object as the ``.type``
attribute will report this exception instead of recursion overflow.
fixes #3485
Mike Bayer [Fri, 17 Jul 2015 15:09:29 +0000 (11:09 -0400)]
- db_opts doesn't need to be a global since we no longer have any global
option for it. keep it local to production of config so values don't
leak in a multi --db situation.