Release Notes: This release features support for Python 3.1 and 3.2, improvements in named cursor efficiency, and better support for large objects and Python/PostgreSQL type conversion (bytearray and memoryview to bytea, hstore support). This release also fixes some bugs related to adaptation of composite types, conversion of negative infinities, and some reference leaks.
Release Notes: This version introduces Python 3 support.
Release Notes: This release introduces many new features related to PostgreSQL 9.0. psycopg now supports two-phase commit (TCP) as specified by the PEP-249, NOTIFY payloads, and the new hstore type.
Release Notes: The major feature of the 2.2 series is the support of both select()-based and "green" coroutine-based asynchronous queries. This release also spots better timezone management, better type-casting, and improved COPY and NOTIFY support.
Release Notes: psycopg 2.0 is now the stable branch.
Release Notes: This release has support for PostgreSQL arrays (mapped to Python list objects) and improved support for BINARY columns (now mapped to Python buffer objects.) It also fixes some outstanding bugs related to GIL, .executemany(), binary and datetime management, and running on 64-bit architectures.
Release Notes: This release implements COPY FROM/COPY TO using the libpq version 3 protocol, implements adapters for all the basic Python types, and adds small, generic fixes.
Release Notes: ZPsycopgDA was added back in this release, using the new connection pooling code. A new DECIMAL typecaster wes added to convert from the PostgreSQL money type to the Python 2.4 Decimal type. Some minor bugs related to fetching query results were fixed.
Release Notes: This release added support for Unicode queries and automatic conversion of PostgreSQL strings to Python Unicode objects. Cursors now support the iterator protocol and have two new methods (.fileno() and .isready()) that make them directly usable in select() calls. A memory corruption bug on Win32 and time rounding errors were also fixed.
Release Notes: This release adds support for time zones when converting to native Python datetime objects. .fetchXXX methods now can return custom row objects instead of plain tuples.