SQLAlchemy 1.3.12 released. SQLAlchemy is a Python SQL database object and kits mapping (ORM) framework. It contains the complete set of enterprise-level persistence patterns, designed for high efficiency and high performance database access.
Updates are as follows:
snake
- Fixed an issue involving lazy = "raise" strategy, in this strategy, ORM delete an object will trigger configured lazy = "raise" simple "use-get" many-style relationship.
- Fixed a regression introduced in 1.3.0, it # 4351 associated with the agent about the reconstruction, the return prevents composite () properties play a role in their associated reference agency.
- Now, is provided on the relationship () and associated signs durability, while also setting viewonly = True, will issue regular warnings, because these signs does not make sense for viewonly = True relationship.
- Fixes the following issues: When the set is assigned to itself as a slice, mutation operation will fail because it will inadvertently erased first set of assignment. Because they do not change the content of the assignment should not generate events, so this operation is now empty operation.
- Fixes an issue: if the transaction "start" in the core engine / connection-level failure, for example due to a network error or due to some transactions recipe database is locked, then get the connection from the connection pool in the context of the conversation, and then immediately return it, even if the connection is not stored in the session state, the ORM session does not close the connection.
sql
- Bug fixes: passed to select () the "distinct" keyword does not like select.distinct () as the string value as "label references"; the contrary, it will increase unconditionally.
- Change the "Unable to resolve label references" exception text to include other types of labels mandatory, namely "DISTINCT" also belongs to PostgreSQL dialect.
sqlite
- Fixed problem solving SQLite as JSON data types assigned "number" relevance behavior, the behavior first was first described in support for SQLite JSON added, the operation returns scalar number JSON values in digital form, rather than be deserialized JSON string.
mssql
- By adding PyODBC level of result handlers, repair support for mssql.DATETIMEOFFSET data type on PyODBC, because it does not contain this data type of native support.
Update Description: https://docs.sqlalchemy.org/en/13/changelog/changelog_13.html#change-1.3.12