🎫`2369` and 🎫`2587` where a nested join with one side
already an aliased select would fail to translate the ON clause on the
outside correctly; in the ORM this could be seen when using a
SELECT statement as a "secondary" table. [ticket:2858]
query being generated would produce a "columns being replaced"
warning for a statement with two same-named columns,
as the internal SELECT wouldn't have use_labels set.
attribute will now honor the "passive" flag
passed to it; as this defaults to ``PASSIVE_OFF``, the function will
by default query the database if the value is not present.
This is a behavioral change vs. 0.8. [ticket:2787]
- Added new method :meth:`.AttributeState.load_history`, works like
:attr:`.AttributeState.history` but also fires loader callables.
with conjunctions, e.g.
``None`` :func:`.expression.null` :func:`.expression.true`
:func:`.expression.false`, including consistency in rendering NULL
in conjunctions, "short-circuiting" of :func:`.and_` and :func:`.or_`
expressions which contain boolean constants, and rendering of
boolean constants and expressions as compared to "1" or "0" for backends
that don't feature ``true``/``false`` constants. [ticket:2804]
specified) is now copied when used in a typed expression, and the
new copy is assigned the actual type of the compared column. Previously,
this logic would occur on the given :func:`.bindparam` in place.
Additionally, a similar process now occurs for :func:`.bindparam` constructs
passed to :meth:`.ValuesBase.values` for a :class:`.Insert` or
:class:`.Update` construct. [ticket:2850]
e.g. values that are normally bound parameters but due to context must
be rendered as strings, typically within DDL constructs such as
CHECK constraints and indexes (note that "literal bind" values
become used by DDL as of 🎫`2742`). A new method
:meth:`.TypeEngine.literal_processor` serves as the base, and
:meth:`.TypeDecorator.process_literal_param` is added to allow wrapping
of a native literal rendering method. [ticket:2838]
- enhance _get_colparams so that we can send flags like literal_binds into
INSERT statements
- add support in PG for inspecting standard_conforming_strings
- add a new series of roundtrip tests based on INSERT of literal plus SELECT
for basic literal rendering in dialect suite
so that the custom type isn't exposed to an operation that is against the
"impl" type's constraint, [ticket:2842]
- this change showed up as some recursion overflow in pickling with labels,
add a __reduce__() there....pickling of expressions is less and less something
that's very viable...
all :attr:`.SchemaItem.info` dictionaries from all :class:`.SchemaItem`
objects within the structure including columns, constraints,
foreign keys, etc. As these dictionaries
are copies, they are independent of the original dictionary.
Previously, only the ``.info`` dictionary of :class:`.Column` was transferred
within this operation, and it was only linked in place, not copied.
[ticket:2716]
of :class:`.ForeignKeyConstraint` is silently ignored on the MySQL
backend, will be reverted as of 0.9; this keyword will now render again, raising
errors on MySQL as it is not understood - the same behavior will also
apply to the ``initially`` keyword. In 0.8, the keywords will remain
ignored but a warning is emitted. Additionally, the ``match`` keyword
now raises a :class:`.CompileError` on 0.9 and emits a warning on 0.8;
this keyword is not only silently ignored by MySQL but also breaks
the ON UPDATE/ON DELETE options.
To use a :class:`.ForeignKeyConstraint`
that does not render or renders differently on MySQL, use a custom
compilation option. An example of this usage has been added to the
documentation, see :ref:`mysql_foreign_keys`.
[ticket:2721] [ticket:2839]
behavior as that of :class:`.mysql.ENUM`. Quotes are not required when
setting up the value, but quotes that are present will be auto-detected
along with a warning. This also helps with Alembic where
the SET type doesn't render with quotes. [ticket:2817]
if it was generated from a :class:`.Column` that didn't specify ``unique``
(where it defaults to ``None``). The flag will now always be ``True`` or
``False``. [ticket:2825]
This enables the subquery eager loader strategy to apply a DISTINCT
to the innermost SELECT subquery, to assist in the case where
duplicate rows are generated by the innermost query which corresponds
to this relationship (there's not yet a general solution to the issue
of dupe rows within subquery eager loading, however, when joins outside
of the innermost subquery produce dupes). When the flag
is set to ``True``, the DISTINCT is rendered unconditionally, and when
it is set to ``None``, DISTINCT is rendered if the innermost relationship
targets columns that do not comprise a full primary key.
The option defaults to False in 0.8 (e.g. off by default in all cases),
None in 0.9 (e.g. automatic by default). Thanks to Alexander Koval
for help with this. [ticket:2836]
mssql to ensure that any literal SQL expression values are
rendered directly as literals, instead of as bound parameters,
within a CREATE INDEX statement. [ticket:2742]
- don't need expression_as_ddl(); literal_binds and include_table
take care of this functionality.
events are called, so that attachment events can be used to dynamically
generate a name for the index based on the parent table and/or
columns. [ticket:2835]
all known cases is provided by :class:`.DefaultDialect`, has been
tightened to expect ``include_columns`` and ``exclude_columns``
arguments without any kw option, reducing ambiguity - previously
``exclude_columns`` was missing. [ticket:2748]
:func:`.foreign` on a :class:`.Column` before association with a parent
:class:`.Table` could produce issues related to the parent table not
rendering within joins, due to the inherent copy operation performed
by an annotation. [ticket:2813]
- The system of loader options has been entirely rearchitected to build
upon a much more comprehensive base, the :class:`.Load` object. This
base allows any common loader option like :func:`.joinedload`,
:func:`.defer`, etc. to be used in a "chained" style for the purpose
of specifying options down a path, such as ``joinedload("foo").subqueryload("bar")``.
The new system supersedes the usage of dot-separated path names,
multiple attributes within options, and the usage of ``_all()`` options.
- Added a new load option :func:`.orm.load_only`. This allows a series
of column names to be specified as loading "only" those attributes,
deferring the rest.
Mike Bayer