]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
12 years agoAdded a conditional import to the ``gaerdbms`` dialect which attempts
Mike Bayer [Sat, 2 Feb 2013 23:40:56 +0000 (18:40 -0500)] 
Added a conditional import to the ``gaerdbms`` dialect which attempts
to import rdbms_apiproxy vs. rdbms_googleapi to work
on both dev and production platforms.  Also now honors the
``instance`` attribute.  Courtesy Sean Lynch.
[ticket:2649]

12 years agochangelog for 0.7
Mike Bayer [Sat, 2 Feb 2013 22:00:54 +0000 (17:00 -0500)] 
changelog for 0.7

12 years agoclean up ordering list docs, [ticket:2557]
Mike Bayer [Sat, 2 Feb 2013 21:46:12 +0000 (16:46 -0500)] 
clean up ordering list docs, [ticket:2557]

12 years agofix markup
Mike Bayer [Sat, 2 Feb 2013 21:20:30 +0000 (16:20 -0500)] 
fix markup

12 years agoThe :meth:`.ColumnOperators.in_` operator will now coerce
Mike Bayer [Sat, 2 Feb 2013 21:17:58 +0000 (16:17 -0500)] 
The :meth:`.ColumnOperators.in_` operator will now coerce
values of ``None`` to :func:`.null`.
[ticket:2496]

12 years agoAdded a new argument to :class:`.Enum` and its base
Mike Bayer [Sat, 2 Feb 2013 01:47:02 +0000 (20:47 -0500)] 
Added a new argument to :class:`.Enum` and its base
:class:`.SchemaType` ``inherit_schema``.  When set to ``True``,
the type will set its ``schema`` attribute of that of the
:class:`.Table` to which it is associated.  This also occurs
during a :meth:`.Table.tometadata` operation; the :class:`.SchemaType`
is now copied in all cases when :meth:`.Table.tometadata` happens,
and if ``inherit_schema=True``, the type will take on the new
schema name passed to the method.   The ``schema`` is important
when used with the Postgresql backend, as the type results in
a ``CREATE TYPE`` statement. [ticket:2657]

12 years agoversion bump
Mike Bayer [Sat, 2 Feb 2013 01:45:29 +0000 (20:45 -0500)] 
version bump

12 years ago- add full docs for like()/ilike()
Mike Bayer [Mon, 28 Jan 2013 18:58:01 +0000 (13:58 -0500)] 
- add full docs for like()/ilike()

12 years agoadd any/all to __all__
Mike Bayer [Mon, 28 Jan 2013 18:57:48 +0000 (13:57 -0500)] 
add any/all to __all__

12 years ago- documentation for any()/all()
Mike Bayer [Mon, 28 Jan 2013 18:49:18 +0000 (13:49 -0500)] 
- documentation for any()/all()

12 years agoAdd ANY/ALL construct support for PostgreSQL's ARRAY type
Audrius Kažukauskas [Mon, 28 Jan 2013 17:58:06 +0000 (19:58 +0200)] 
Add ANY/ALL construct support for PostgreSQL's ARRAY type

12 years ago- use SQL constructs here for databases that need to escape names like "data"
Mike Bayer [Sun, 27 Jan 2013 22:15:25 +0000 (17:15 -0500)] 
- use SQL constructs here for databases that need to escape names like "data"

12 years agoFixed bug where :meth:`.Table.tometadata` would fail if a
Mike Bayer [Sun, 27 Jan 2013 16:23:59 +0000 (11:23 -0500)] 
Fixed bug where :meth:`.Table.tometadata` would fail if a
:class:`.Column` had both a foreign key as well as an
alternate ".key" name for the column.   Also in 0.7.10.
[ticket:2643]

12 years agoFixing a code block in the PostgreSQL base dialect docstring.
Taavi Burns [Sat, 26 Jan 2013 03:48:34 +0000 (22:48 -0500)] 
Fixing a code block in the PostgreSQL base dialect docstring.

12 years ago#2629
Mike Bayer [Fri, 25 Jan 2013 17:57:59 +0000 (12:57 -0500)] 
#2629

insert().returning() raises an informative CompileError if attempted
to compile on a dialect that doesn't support RETURNING.

12 years agothe consideration of a pending object as
Mike Bayer [Fri, 25 Jan 2013 02:31:23 +0000 (21:31 -0500)] 
the consideration of a pending object as
an "orphan" has been modified to more closely match the
behavior as that of persistent objects, which is that the object
is expunged from the :class:`.Session` as soon as it is
de-associated from any of its orphan-enabled parents.  Previously,
the pending object would be expunged only if de-associated
from all of its orphan-enabled parents.  The new flag ``legacy_is_orphan``
is added to :func:`.orm.mapper` which re-establishes the
legacy behavior. [ticket:2655]

12 years agoFixed the (most likely never used) "@collection.link" collection
Mike Bayer [Mon, 21 Jan 2013 23:17:10 +0000 (18:17 -0500)] 
Fixed the (most likely never used) "@collection.link" collection
method, which fires off each time the collection is associated
or de-associated with a mapped object - the decorator
was not tested or functional.  The decorator method
is now named :meth:`.collection.linker` though the name "link"
remains for backwards compatibility.  Courtesy Luca Wehrstedt.
[ticket:2653]

12 years agoMerged in lerks/sqlalchemy (pull request #36)
Mike Bayer [Mon, 21 Jan 2013 22:59:14 +0000 (17:59 -0500)] 
Merged in lerks/sqlalchemy (pull request #36)

Fix the collection.link decorator

12 years ago- Made some fixes to the system of producing custom instrumented
Mike Bayer [Mon, 21 Jan 2013 22:57:24 +0000 (17:57 -0500)] 
- Made some fixes to the system of producing custom instrumented
collections, mainly that the usage of the @collection decorators
will now honor the __mro__ of the given class, applying the
logic of the sub-most classes' version of a particular collection
method.   Previously, it wasn't predictable when subclassing
an existing instrumented class such as :class:`.MappedCollection`
whether or not custom methods would resolve correctly.
[ticket:2654]

- The undocumented (and hopefully unused) system of producing
custom collections using an ``__instrumentation__`` datastructure
associated with the collection has been removed, as this was a complex
and untested feature which was also essentially redundant versus the
decorator approach.   Other internal simplifcations to the
orm.collections module have been made as well.

12 years agoclean all pep8 warnings
Mike Bayer [Mon, 21 Jan 2013 21:16:57 +0000 (16:16 -0500)] 
clean all pep8 warnings

12 years ago- replace mssql_ordering with generalized #695 solution
Mike Bayer [Sun, 20 Jan 2013 17:43:39 +0000 (12:43 -0500)] 
- replace mssql_ordering with generalized #695 solution
- documentation for mssql index options plus changelog and fixes

12 years agoMerged in dharland/sqlalchemy (pull request #35)
Mike Bayer [Sun, 20 Jan 2013 17:23:39 +0000 (12:23 -0500)] 
Merged in dharland/sqlalchemy (pull request #35)

Add extra mssql dialect options to Index

12 years agoRemoves an errant space character that caused a newline break in the sphinx output.
Taavi Burns [Thu, 17 Jan 2013 22:31:07 +0000 (17:31 -0500)] 
Removes an errant space character that caused a newline break in the sphinx output.

12 years agofix incorrect quoting in mysql indexes
Mike Bayer [Thu, 17 Jan 2013 15:19:54 +0000 (10:19 -0500)] 
fix incorrect quoting in mysql indexes

12 years ago- fix a format issue in the create index
Mike Bayer [Thu, 17 Jan 2013 15:13:08 +0000 (10:13 -0500)] 
- fix a format issue in the create index

12 years agoMutableDict.__delitem__ should require only 'key' argument
Audrius Kažukauskas [Thu, 17 Jan 2013 10:35:51 +0000 (12:35 +0200)] 
MutableDict.__delitem__ should require only 'key' argument

12 years ago- "postgres_where" has been removed
Mike Bayer [Thu, 17 Jan 2013 04:53:00 +0000 (23:53 -0500)] 
- "postgres_where" has been removed

12 years ago:class:`.Index` now supports arbitrary SQL expressions and/or
Mike Bayer [Thu, 17 Jan 2013 02:04:32 +0000 (21:04 -0500)] 
:class:`.Index` now supports arbitrary SQL expressions and/or
functions, in addition to straight columns.   Common modifiers
include using ``somecolumn.desc()`` for a descending index and
``func.lower(somecolumn)`` for a case-insensitive index, depending on the
capabilities of the target backend.
[ticket:695]

12 years ago- preparing for #695, modernize constraint/index tests
Mike Bayer [Thu, 17 Jan 2013 01:10:55 +0000 (20:10 -0500)] 
- preparing for #695, modernize constraint/index tests

12 years agoFix the collection.link decorator
Luca Wehrstedt [Wed, 16 Jan 2013 15:26:15 +0000 (16:26 +0100)] 
Fix the collection.link decorator

12 years agoAdd mssql_include option for mssql dialect
Derek Harland [Mon, 14 Jan 2013 04:03:38 +0000 (17:03 +1300)] 
Add mssql_include option for mssql dialect

12 years agoAdd mssql_ordering option for mssql dialect
Derek Harland [Mon, 14 Jan 2013 04:02:20 +0000 (17:02 +1300)] 
Add mssql_ordering option for mssql dialect

12 years agoAdd mssql_clustered option for mssql dialect
Derek Harland [Mon, 14 Jan 2013 04:00:37 +0000 (17:00 +1300)] 
Add mssql_clustered option for mssql dialect

12 years agoAdded a py3K conditional around unnecessary .decode()
Mike Bayer [Sun, 13 Jan 2013 00:51:13 +0000 (19:51 -0500)] 
Added a py3K conditional around unnecessary .decode()
call in mssql information schema, fixes reflection
in Py3K. Also in 0.7.10. [ticket:2638]

12 years agoremove all specifics from the "supported features" section as this
Mike Bayer [Sat, 12 Jan 2013 22:43:12 +0000 (17:43 -0500)] 
remove all specifics from the "supported features" section as this
is not maintainable

12 years ago- update some profiling stats due to the #2650 change
Mike Bayer [Sat, 12 Jan 2013 22:36:07 +0000 (17:36 -0500)] 
- update some profiling stats due to the #2650 change

12 years ago- add workaround for sqlite memusage tests, so no longer need to count to 220/skip...
Mike Bayer [Sat, 12 Jan 2013 22:21:35 +0000 (17:21 -0500)] 
- add workaround for sqlite memusage tests, so no longer need to count to 220/skip tests
- Fixed potential memory leak which could occur if an
arbitrary number of :class:`.sessionmaker` objects
were created.   The anonymous subclass created by
the sessionmaker, when dereferenced, would not be garbage
collected due to remaining class-level references from the
event package.  This issue also applies to any custom system
that made use of ad-hoc subclasses in conjunction with
an event dispatcher.  Also in 0.7.10. [ticket:2650]

12 years ago- changelog for pullreq 32
Mike Bayer [Sat, 12 Jan 2013 17:14:15 +0000 (12:14 -0500)] 
- changelog for pullreq 32
- Fixed a regression whereby the "collation" parameter
of the character types CHAR, NCHAR, etc. stopped working,
as "collation" is now supported by the base string types.
The TEXT, NCHAR, CHAR, VARCHAR types within the
MSSQL dialect are now synonyms for the base types.
- move out the type rendering tests into DB-agnostic tests
and remove some of the old "create" statements.  tests here
are still very disorganized.

12 years agoMerged in dharland/sqlalchemy (pull request #32: Allow the MSSQL dialect to support...
Mike Bayer [Sat, 12 Jan 2013 16:43:36 +0000 (11:43 -0500)] 
Merged in dharland/sqlalchemy (pull request #32: Allow the MSSQL dialect to support identity columns that are not part of the primary key)

12 years agonote the change needed to the beaker example also
Mike Bayer [Tue, 8 Jan 2013 16:05:13 +0000 (11:05 -0500)] 
note the change needed to the beaker example also

12 years agoFixed a regression in the examples/dogpile_caching example
Mike Bayer [Tue, 8 Jan 2013 15:59:26 +0000 (10:59 -0500)] 
Fixed a regression in the examples/dogpile_caching example
which was due to the change in #2614.

12 years ago:meth:`.Query.merge_result` can now load rows from an outer join
Mike Bayer [Tue, 8 Jan 2013 15:55:46 +0000 (10:55 -0500)] 
:meth:`.Query.merge_result` can now load rows from an outer join
where an entity may be ``None`` without throwing an error.
[ticket:2640]

12 years agoTweaked the "REQUIRED" symbol used by the compiler to identify
Mike Bayer [Tue, 8 Jan 2013 15:04:55 +0000 (10:04 -0500)] 
Tweaked the "REQUIRED" symbol used by the compiler to identify
INSERT/UPDATE bound parameters that need to be passed, so that
it's more easily identifiable when writing custom bind-handling
code. [ticket:2648]

12 years ago- remove the test against bindparam('x') for the bind/result SQL expressions
Mike Bayer [Wed, 2 Jan 2013 16:37:20 +0000 (11:37 -0500)] 
- remove the test against bindparam('x') for the bind/result SQL expressions
- apply a consistent approach to the "see if method X is implemented" conditionals

12 years agoFixed bug in :func:`.postgresql.array` construct whereby using it
Mike Bayer [Wed, 2 Jan 2013 16:26:37 +0000 (11:26 -0500)] 
Fixed bug in :func:`.postgresql.array` construct whereby using it
inside of an :func:`.expression.insert` construct would produce an
error regarding a parameter issue in the ``self_group()`` method.

12 years agohappy new year (see #2645)
Diana Clarke [Wed, 2 Jan 2013 00:31:00 +0000 (19:31 -0500)] 
happy new year (see #2645)

12 years agohappy new year (see #2645)
Diana Clarke [Wed, 2 Jan 2013 00:08:19 +0000 (19:08 -0500)] 
happy new year (see #2645)

12 years agohappy new year (see #2645)
Diana Clarke [Tue, 1 Jan 2013 18:46:21 +0000 (13:46 -0500)] 
happy new year (see #2645)

12 years agoRemoves garbage from the end of an example SQL query line.
Taavi Burns [Mon, 31 Dec 2012 05:32:01 +0000 (00:32 -0500)] 
Removes garbage from the end of an example SQL query line.

12 years agoAdjusts example code so it can be successfully pasted into a REPL.
Taavi Burns [Mon, 31 Dec 2012 05:31:43 +0000 (00:31 -0500)] 
Adjusts example code so it can be successfully pasted into a REPL.

12 years agoFixes grammar in docstring.
Taavi Burns [Mon, 31 Dec 2012 05:21:05 +0000 (00:21 -0500)] 
Fixes grammar in docstring.

12 years agoExtended the :doc:`/core/inspection` system so that all Python descriptors
Mike Bayer [Sun, 30 Dec 2012 00:31:28 +0000 (19:31 -0500)] 
Extended the :doc:`/core/inspection` system so that all Python descriptors
associated with the ORM or its extensions can be retrieved.
This fulfills the common request of being able to inspect
all :class:`.QueryableAttribute` descriptors in addition to
extension types such as :class:`.hybrid_property` and
:class:`.AssociationProxy`.  See :attr:`.Mapper.all_orm_descriptors`.

12 years agochangelog for pullreq 33
Mike Bayer [Sat, 29 Dec 2012 01:45:39 +0000 (20:45 -0500)] 
changelog for pullreq 33

12 years agoFixes issue where GAE error handling can cause AttributeError: 'NoneType' object...
Owen Nelson [Sat, 29 Dec 2012 01:20:23 +0000 (20:20 -0500)] 
Fixes issue where GAE error handling can cause AttributeError: 'NoneType' object has no attribute 'group'

12 years agoGAE dialect now supports the use of credentials
Owen Nelson [Sat, 29 Dec 2012 01:18:20 +0000 (20:18 -0500)] 
GAE dialect now supports the use of credentials

12 years agofixing typos in the types docs (see #2639)
Diana Clarke [Sun, 23 Dec 2012 23:15:12 +0000 (18:15 -0500)] 
fixing typos in the types docs (see #2639)

12 years agotypo in joined table inheritance docs see #2641
Diana Clarke [Sun, 23 Dec 2012 21:13:59 +0000 (16:13 -0500)] 
typo in joined table inheritance docs see #2641

12 years ago- add tests to ensure SELECT of dynamic collection not emitted
Mike Bayer [Sat, 22 Dec 2012 23:40:06 +0000 (18:40 -0500)] 
- add tests to ensure SELECT of dynamic collection not emitted
on regular append/remove history
- fix the real cause of the original #2637 issue, backrefs call upon
the "pop()" method now which wasn't implemented for Dynamic

12 years ago- revert the full iteration of the collection for a passive history event; that's
Mike Bayer [Sat, 22 Dec 2012 15:36:55 +0000 (10:36 -0500)] 
- revert the full iteration of the collection for a passive history event; that's
the wrong behavior, and for the original #2637 issue we will have to fix the
association proxy, which is #2642

12 years ago- significantly rework the approach to collection events and history within DynamicAt...
Mike Bayer [Fri, 21 Dec 2012 22:53:57 +0000 (17:53 -0500)] 
- significantly rework the approach to collection events and history within DynamicAttributeImpl
- Fixes to the "dynamic" loader on :func:`.relationship`, includes
that backrefs will work properly even when autoflush is disabled,
history events are more accurate in scenarios where multiple add/remove
of the same object occurs, as can often be the case in conjunction
with the association proxy.  [ticket:2637]

12 years ago- cleanup on these tests
Mike Bayer [Fri, 21 Dec 2012 18:29:48 +0000 (13:29 -0500)] 
- cleanup on these tests
- add "extend()" to AppenderQuery

12 years agodocs - missing 'attr' in AssociationProxy join see #2636
Diana Clarke [Sun, 16 Dec 2012 13:01:08 +0000 (08:01 -0500)] 
docs - missing 'attr' in AssociationProxy join see #2636

12 years agoAdded tag rel_0_8_8b2 for changeset 332560b1fd0917a0e7dbeb295d48d045ee6f6887
Mike Bayer [Fri, 14 Dec 2012 20:52:30 +0000 (15:52 -0500)] 
Added tag rel_0_8_8b2 for changeset 332560b1fd0917a0e7dbeb295d48d045ee6f6887

12 years ago0.8.0b2 rel_0_8_0b2
Mike Bayer [Fri, 14 Dec 2012 20:40:05 +0000 (15:40 -0500)] 
0.8.0b2

12 years agothis comment is entirely from some ancient version of the code
Mike Bayer [Fri, 14 Dec 2012 15:33:23 +0000 (10:33 -0500)] 
this comment is entirely from some ancient version of the code

12 years agoMore adjustment to this SQLite related issue which was released in
Mike Bayer [Fri, 14 Dec 2012 15:29:46 +0000 (10:29 -0500)] 
More adjustment to this SQLite related issue which was released in
0.7.9, to intercept legacy SQLite quoting characters when reflecting
foreign keys.  In addition to intercepting double quotes, other
quoting characters such as brackets, backticks, and single quotes
are now also intercepted. [ticket:2568]

12 years agoThe :meth:`.Query.select_from` method can now be used with a
Mike Bayer [Thu, 13 Dec 2012 23:45:15 +0000 (18:45 -0500)] 
The :meth:`.Query.select_from` method can now be used with a
:func:`.aliased` construct without it interfering with the entities
being selected.  [ticket:2635]

12 years agoFixed a regression caused by :ticket:`2410` whereby a
Mike Bayer [Tue, 11 Dec 2012 21:31:41 +0000 (16:31 -0500)] 
Fixed a regression caused by :ticket:`2410` whereby a
:class:`.CheckConstraint` would apply itself back to the
original table during a :meth:`.Table.tometadata` operation, as
it would parse the SQL expression for a parent table. The
operation now copies the given expression to correspond to the
new table.
[ticket:2633]

12 years agoadd future
Mike Bayer [Tue, 11 Dec 2012 15:34:33 +0000 (10:34 -0500)] 
add future

12 years agoFixed table reflection for Oracle when accessing a synonym that refers
Mike Bayer [Sun, 9 Dec 2012 23:28:08 +0000 (18:28 -0500)] 
Fixed table reflection for Oracle when accessing a synonym that refers
to a DBLINK remote database; while the syntax has been present in the
Oracle dialect for some time, up until now it has never been tested.
The syntax has been tested against a sample database linking to itself,
however there's still some uncertainty as to what should be used for the
"owner" when querying the remote database for table information.
Currently, the value of "username" from user_db_links is used to
match the "owner".  [ticket:2619]

12 years agotypo
Mike Bayer [Sun, 9 Dec 2012 19:43:49 +0000 (14:43 -0500)] 
typo

12 years ago- documentation and changelog for [ticket:2601]
Mike Bayer [Sun, 9 Dec 2012 19:12:22 +0000 (14:12 -0500)] 
- documentation and changelog for [ticket:2601]

12 years agoadd a doc for the [ticket:2631] quirk
Mike Bayer [Sun, 9 Dec 2012 01:39:32 +0000 (20:39 -0500)] 
add a doc for the [ticket:2631] quirk

12 years agoFixed bug where using server_onupdate=<FetchedValue|DefaultClause>
Mike Bayer [Sun, 9 Dec 2012 01:28:43 +0000 (20:28 -0500)] 
Fixed bug where using server_onupdate=<FetchedValue|DefaultClause>
without passing the "for_update=True" flag would apply the default
object to the server_default, blowing away whatever was there.
The explicit for_update=True argument shouldn't be needed with this usage
(especially since the documentation shows an example without it being
used) so it is now arranged internally using a copy of the given default
object, if the flag isn't set to what corresponds to that argument.
Also in 0.7.10. [ticket:2631]

12 years agoFixed bug whereby using a label_length on dialect that was smaller
Mike Bayer [Sun, 9 Dec 2012 00:25:04 +0000 (19:25 -0500)] 
Fixed bug whereby using a label_length on dialect that was smaller
than the size of actual column identifiers would fail to render
the columns correctly in a SELECT statement.
[ticket:2610]

12 years ago- refactor test_labels into all compiler/defaultdialect tests
Mike Bayer [Sun, 9 Dec 2012 00:10:44 +0000 (19:10 -0500)] 
- refactor test_labels into all compiler/defaultdialect tests

12 years agoattempt to hide these from 2to3
Mike Bayer [Sat, 8 Dec 2012 22:50:51 +0000 (17:50 -0500)] 
attempt to hide these from 2to3

12 years ago- version check for sqlite on multivalues is 3.7.11
Mike Bayer [Sat, 8 Dec 2012 21:18:21 +0000 (16:18 -0500)] 
- version check for sqlite on multivalues is 3.7.11

12 years agointernally at least refer to multirow as "multivalues", to distinguish between
Mike Bayer [Sat, 8 Dec 2012 21:17:20 +0000 (16:17 -0500)] 
internally at least refer to multirow as "multivalues", to distinguish between
an INSERT that's used in executemany() as opposed to one which has a VALUES
clause with multiple entries.

12 years ago- multivalued inserts, [ticket:2623]
Mike Bayer [Sat, 8 Dec 2012 19:25:42 +0000 (14:25 -0500)] 
- multivalued inserts, [ticket:2623]

- update "not supported" messages for empty inserts, mutlivalue inserts

- rework the ValuesBase approach for multiple value sets so that stmt.parameters
does store a list for multiple values; the _has_multiple_parameters flag now indicates
which of the two modes the statement is within.  it now raises exceptions if a subsequent
call to values() attempts to call a ValuesBase with one mode in the style of the other
mode; that is, you can't switch a single- or multi- valued ValuesBase to the other mode,
and also if a multiple value is passed simultaneously with a kwargs set.
Added tests for these error conditions

- Calling values() multiple times in multivalue mode now extends the parameter list to
include the new parameter sets.

- add error/test if multiple *args were passed to ValuesBase.values()

- rework the compiler approach for multivalue inserts, back to where
_get_colparams() returns the same list of (column, value) as before, thereby
maintaining the identical number of append() and other calls when multivalue
is not enabled.  In the case of multivalue, it makes a last-minute switch to return
a list of lists instead of the single list.  As it constructs the additional lists, the inline
defaults and other calculated default parameters of the first parameter
set are copied into the newly generated lists so that these features continue
to function for a multivalue insert.   Multivalue inserts now add no additional
function calls to the compilation for regular insert constructs.

- parameter lists for multivalue inserts now includes an integer index for all
parameter sets.

- add detailed documentation for ValuesBase.values(), including careful wording
to describe the difference between multiple values and an executemany() call.

- add a test for multivalue insert + returning - it works !

- remove the very old/never used "postgresql_returning"/"firebird_returning" flags.

12 years agomerge latest default
Mike Bayer [Sat, 8 Dec 2012 16:23:21 +0000 (11:23 -0500)] 
merge latest default

12 years agofixing InstrumentationManager links
Diana Clarke [Sat, 8 Dec 2012 03:42:34 +0000 (22:42 -0500)] 
fixing InstrumentationManager links

12 years agofixing sessionmaker link
Diana Clarke [Sat, 8 Dec 2012 03:23:25 +0000 (22:23 -0500)] 
fixing sessionmaker link

12 years agoThe Oracle LONG type, while an unbounded text type, does not appear
Mike Bayer [Fri, 7 Dec 2012 00:30:49 +0000 (19:30 -0500)] 
The Oracle LONG type, while an unbounded text type, does not appear
to use the cx_Oracle.LOB type when result rows are returned,
so the dialect has been repaired to exclude LONG from
having cx_Oracle.LOB filtering applied.  Also in 0.7.10.
[ticket:2620]

12 years agomerge dance
Mike Bayer [Fri, 7 Dec 2012 00:14:39 +0000 (19:14 -0500)] 
merge dance

12 years agoRepaired the usage of ``.prepare()`` in conjunction with
Mike Bayer [Fri, 7 Dec 2012 00:10:06 +0000 (19:10 -0500)] 
Repaired the usage of ``.prepare()`` in conjunction with
cx_Oracle so that a return value of ``False`` will result
in no call to ``connection.commit()``, hence avoiding
"no transaction" errors.   Two-phase transactions have
now been shown to work in a rudimental fashion with
SQLAlchemy and cx_oracle, however are subject to caveats
observed with the driver; check the documentation
for details.  Also in 0.7.10.
[ticket:2611]

12 years agofixing broken links (see #2625)
Diana Clarke [Thu, 6 Dec 2012 05:00:17 +0000 (00:00 -0500)] 
fixing broken links (see #2625)

12 years agofixing broken links (see #2625)
Diana Clarke [Thu, 6 Dec 2012 04:19:24 +0000 (23:19 -0500)] 
fixing broken links (see #2625)

12 years agocompiler: add support for multirow inserts
Idan Kamara [Wed, 5 Dec 2012 22:11:52 +0000 (00:11 +0200)] 
compiler: add support for multirow inserts

Some databases support this syntax for inserts:

INSERT INTO table (id, name) VALUES
('v1', 'v2'),
('v3', 'v4');

which greatly increases INSERT speed.

It is now possible to pass a list of lists/tuples/dictionaries as
the values param to the Insert construct. We convert it to a flat
dictionary so we can continue using bind params. The above query
will be converted to:

INSERT INTO table (id, name) VALUES
(:id, :name),
(:id0, :name0);

Currently only supported on postgresql, mysql and sqlite.

12 years agocompiler: adjust _get_colparams to return the columns and parameters in separate...
Idan Kamara [Wed, 5 Dec 2012 21:45:49 +0000 (23:45 +0200)] 
compiler: adjust _get_colparams to return the columns and parameters in separate lists

12 years agoupdatebase: add support for multi parameters
Idan Kamara [Wed, 5 Dec 2012 22:12:16 +0000 (00:12 +0200)] 
updatebase: add support for multi parameters

12 years agoAdd several MSSQL dialect tests for sequence declarations
Derek Harland [Tue, 4 Dec 2012 01:49:03 +0000 (14:49 +1300)] 
Add several MSSQL dialect tests for sequence declarations

12 years agoMSSQL Dialect should set sequence columns to be non null
Derek Harland [Tue, 4 Dec 2012 01:48:37 +0000 (14:48 +1300)] 
MSSQL Dialect should set sequence columns to be non null

12 years agoMerge changes from official sqlalchemy repo
Derek Harland [Tue, 4 Dec 2012 01:24:59 +0000 (14:24 +1300)] 
Merge changes from official sqlalchemy repo

12 years agochangelog
Mike Bayer [Tue, 4 Dec 2012 00:59:05 +0000 (19:59 -0500)] 
changelog

12 years agomerge decimal thing
Mike Bayer [Tue, 4 Dec 2012 00:57:28 +0000 (19:57 -0500)] 
merge decimal thing

12 years agoAllow the MSSQL dialect to support identity columns that are not part of the primary key
Derek Harland [Tue, 4 Dec 2012 00:56:48 +0000 (13:56 +1300)] 
Allow the MSSQL dialect to support identity columns that are not part of the primary key

12 years agoThe :class:`.MutableComposite` type did not allow for the
Mike Bayer [Tue, 4 Dec 2012 00:49:42 +0000 (19:49 -0500)] 
The :class:`.MutableComposite` type did not allow for the
:meth:`.MutableBase.coerce` method to be used, even though
the code seemed to indicate this intent, so this now works
and a brief example is added.  As a side-effect,
the mechanics of this event handler have been changed so that
new :class:`.MutableComposite` types no longer add per-type
global event handlers.  Also in 0.7.10

[ticket:2624]

12 years agowhitespace
Mike Bayer [Tue, 4 Dec 2012 00:49:05 +0000 (19:49 -0500)] 
whitespace

12 years agocleanup
Mike Bayer [Tue, 4 Dec 2012 00:01:25 +0000 (19:01 -0500)] 
cleanup