Turbodbc 2.5.0

Highlights:

• Support for Apache Arrow 0.8.x
• fetchallarrow() optionally supports data-dependent integer types to save space when possible

Please see the Changelog for more details.

Turbodbc 2.4.1

Highlights:

• Fixed a memory leak in the Cursor.fetchallarrow()

Please see the Changelog for more details.

Turbodbc 2.4.0

Highlights:

• Cursor.executemanycolumns() supports Apache Arrow pyarrow.Table, allowing convenient inserts from pandas.DataFrame

Please see the Changelog for more details

Turbodbc 2.3.0

Highlights:

• cursor.fetchallarrow() now supports representing string fields as Apache Arrow's dictionary arrays. This is helpful for string fields containing only a limited number of distinct entries, e.g., cities, countries, colors, etc.

Please see the changelog for more details

Turbodbc 2.2.0

Highlights:

• Large decimals with more than 18 digits can be retrieved as 64-bit data types
• Size of large VARCHAR fields can be limited
• executemanycolumns() supports datetime64[ns] for better pandas integration

Please see the changelog for more details.

Turbodbc 2.1.0

Highlight:

• New cursor method executemanycolumns() accepts NumPy columns for extra-speedy inserts without intermediate conversion of numbers, etc. into Python objects.

For more details, please have a look at the changelog

Turbodbc 2.0.0

Highlight:

• New built-in Apache Arrow support for retrieving result sets via cursor.fetchallarrow()

For more details, please have a look at the changelog

Turbodbc 1.1.2

Highlight:

• Added options to control automatic commit behavior

For more details, please have a look at the changelog

Turbodbc 1.1.1

Highlights:

• Official Windows support (64 bit, Python 3.5 and 3.6) including automatic testing and upload of binary wheels
• New cursor.fetchnumpybatches() method allows to iterate over batches of NumPy arrays

For more details, please have a look at the changelog