1. idank
  2. sqlalchemy

Commits

Author Commit Message Date Builds
Mike Bayer
seems like the recent itertools add to select()._get_display_froms() adds overhead in 2.4? not sure why
Mike Bayer
fix up some unit tests
Mike Bayer
- merge() may actually work now, though we've heard that before... - merge() uses the priamry key attributes on the object if _instance_key not present. so merging works for instances that dont have an instnace_key, will still issue UPDATE for existing rows. - improved collection behavior for merge() - will remove elements from a destination collection that are not in the source. - fixed naive set-mutation issue in Select._get_display_froms - simplified fixtures…
jason kirtland
- Tighten up r4399 _set_iterable docs
jason kirtland
- Light collections refactor, added public collections.bulk_replace. - Collection attribs gain some private load-from-iterable flexiblity.
Mike Bayer
weird, old cruft
Mike Bayer
- removed redundant get_history() method - the little bit at the bottom of _sort_circular_dependencies is absolutely covered by test/orm/cycles.py ! removing it breaks the test as run on PG.
jason kirtland
C-u 66 C-x f M-q
ram
MSSQL adjustments to pyodbc connection string building
ram
Add a new 'driver' keyword to the MSSQL pyodbc Dialect. Refresh items that were recently reverted by another checkin
Mike Bayer
- reverted previous "strings instead of tuples" change due to more specific test results showing tuples faster - changed cache decorator call on default_schema_name call to a connection.info specific one
jason kirtland
*whistle*
jason kirtland
- Removed cache decorator.
Mike Bayer
some cache decorator calls...
Mike Bayer
using concatenated strings as keys in generated_ids collection; they hash slightly faster than tuples
Mike Bayer
- schema-qualified tables now will place the schemaname ahead of the tablename in all column expressions as well as when generating column labels. This prevents cross- schema name collisions in all cases [ticket:999] - the "use_schema" argument to compiler.visit_column() is removed. It uses schema in all cases now. - added a new test to the PG dialect to test roundtrip insert/update/delete/select statements with full schema qualification
ram
MSSQL fixes for tickets 979, 916, 884
Mike Bayer
- added _from_self() - changelog authoring
Mike Bayer
- rearranged LoaderStrategies a bit - removed awareness of "dynamic" from attributes and replaced with "impl_class" - moved DynaLoader into dynamic.py - removed create_strategy() method from StrategizedProperty; they set up 'strategy_class' so that StrategizedProperty treats the default the same as the optional loaders
Mike Bayer
turned starargs conversion to a decorator, per jek's advice. select().order_by()/group_by() already take *args.
jason kirtland
- Added PendingDeprecationWarning support - Deprecation decorator is now a real decorator
Mike Bayer
- declarative_base() takes optional kwarg "mapper", which is any callable/class/method that produces a mapper, such as declarative_base(mapper=scopedsession.mapper). This property can also be set on individual declarative classes using the "__mapper_cls__" property.
Mike Bayer
- merged with_polymorphic branch, which was merged with query_columns branch - removes everything to do with select_table, which remains as a keyword argument synonymous with with_polymorphic=('*', select_table). - all "polymorphic" selectables find their way to Query by way of _set_select_from() now, so that all joins/aliasing/eager loads/etc. is handled consistently. Mapper has methods for producing polymorphic selectables so that Query and eagerloaders alike can get to…
Mike Bayer
- can now allow selects which correlate all FROM clauses and have no FROM themselves. These are typically used in a scalar context, i.e. SELECT x, (SELECT x WHERE y) FROM table. Requires explicit correlate() call.
jason kirtland
- Notes for r4338
Mike Bayer
- fixed SQL function truncation of trailing underscores [ticket:996]
jason kirtland
- Added generic func.random (non-standard SQL)
Mike Bayer
a few more tweaks
Mike Bayer
removed AbstractClauseProcessor, merged its copy-and-visit behavior into ClauseVisitor
Mike Bayer
- already-compiled mappers will still trigger compiles of other uncompiled mappers when used [ticket:995]
  1. Prev
  2. Next