What's New

Adds support for PyArrow 21.0

Adds support for conversion of unsupported Arrow data types in write functions with auto_convert parameters

Makes pandas an optional dependency

Adds support for free-threaded Python 3.13 on Windows

Adds support for building against libbson 2.0

Adds support for strict type adherence for data if a schema is provided. If a field value contains a type mismatch with the schema's type for that field, a TypeError will be raised. Note that NaN is a valid type for all fields. To suppress these errors and instead silently convert such mismatches to NaN, pass the allow_invalid=True argument to your pymongoarrow API call.

Adds support for PyArrow 20.0

Adds support for free-threaded Python on Linux and macOS

Adds support for PyArrow 19.0

Adds support for the pyarrow.Decimal128 data type

Adds support for PyArrow 18.0 and Python 3.13

Fixes an issue that caused data loss on documents with nested data containing null or extension types

Adds support for PyArrow's null type

Checks the status of all PyArrow C API calls to prevent segmentation faults

Adds support for PyArrow 17.0

Adds support for nested ObjectId values in Polars conversions

Added support for PyArrow 16.0.

Added a top-level Contributing guide.

Added an optional bool flag to the write() function to skip writing null fields.

Migrated documentation to the MongoDB Docs site.

Support for Polars

Support for PyArrow.DataTypes: large_list, large_string, date32, date64

Support for PyArrow 14.0.

Support for Python 3.12.

Support for PyArrow 13.0.

Revert bug fix for nested extension objects in auto schema, since it caused a performance regression.

Bug fix for projection on nested fields.

Bug fix for nested extension objects in auto schema.

Support BSON binary type.

Support BSON Decimal128 type.

Support Pandas 2.0 and Pandas extension types.

Support PyArrow 12.0.

Added support for BSON Embedded Document type.

Added support for BSON Array type.

Support PyArrow 11.0.

Added wheels for Linux AArch64 and Python 3.11.

Fixed handling of time zones in schema auto-discovery.

Fixed ImportError on Windows by building libbson in "Release" mode.

Support PyArrow 10.0.

Fixed auto-discovery of schemas for aggregation and numpy methods.

Added documentation for auto-discovery of schemas.

Support auto-discovery of schemas in find/aggregate_*_all methods. If the schema is not given, it will be inferred using the first document in the result set.

Support PyArrow 9.0.

Improve error message for lib ImportError.

Support for Decimal128 type.

Support for macOS arm64 architecture on Python 3.9+.

Support for writing tabular datasets (materialized as PyArrow Tables, Pandas DataFrames, or NumPy arrays) to MongoDB by using the write() function. For more information, see the Quick Start guide for more info.

Support for PyArrow 7.0.

Support for the ObjectId type.

Improve error message when schema contains an unsupported type.

Add support for BSON string type.

Add support for BSON boolean type.

Upgraded to bundle libbson 1.21.1. If installing from source, the minimum supported libbson version is now 1.21.0.

Dropped Python 3.6 support (it was dropped in PyArrow 7.0).

Support for PyMongo 4.0.

Support for Python 3.10.

Support for Windows.

The find_arrow_all() method now accepts a user-provided projection.

The find_arrow_all() method now accepts a session object.

Note: PyMongoArrow now requires pyarrow v6.0.x.

Fixed a bug that caused Linux wheels to be created without the appropriate manylinux platform tags.

Support for efficiently converting find and aggregate query result sets into Arrow/Pandas/Numpy data structures.

Support for patching PyMongo's APIs by using the monkey.patch_all() method.

Support for loading the following BSON types:

• 64-bit binary floating point

• 32-bit integer

• 64-bit integer

• Timestamp