Release notes#

PyStarburst 0.13.0 - June 19, 2026#

Features#

  • Add methods DataFrame.to_arrow_batches() and DataFrame.to_arrow_table()

Fixes#

  • Fix DataFrame.to_local_iterator() truncating results at the first protocol response page. The iterator now correctly paginates through all rows and runs on its own cursor, isolated from subsequent actions on the same DataFrame.

PyStarburst 0.12.1 - May 29, 2026#

Dependencies#

  • Restrict pydantic to >=2.12.5,<2.13, as 2.13.4 causes pystarburst to crash on import

PyStarburst 0.12.0 - April 29, 2026#

Features#

  • Add Python 3.14 support

  • Add Arrow spooling support for DataFrame.to_pandas()

Under the Hood#

  • Migrate to Pydantic V2

PyStarburst 0.11.0 - February 05, 2026#

Under the Hood#

  • Drop support for end-of-life Python 3.9

Dependencies#

  • Bump dependencies to address CVEs

PyStarburst 0.10.0 - August 06, 2025#

Features#

  • Add support for Python 3.13

Dependencies#

  • Bump trino-python-client to 0.335.0

PyStarburst 0.9.0 - July 08, 2024#

Features#

  • Add session.cache_results method that allows caching dataframes

  • Add df.write.copy_into_location method that allows export dataframe to external location

Dependencies#

  • Bump pydantic to v2 and adjust code to still use pydantic v1

  • Bump dependencies and trino-python-client to 0.329.0

PyStarburst 0.8.0 - April 30, 2024#

Breaking Changes#

  • Use JodaTime’s DateTimeFormat. Previously we used Teradata functions underneath,
    which for date and time year-month-day hour:minute:second use format:
    yyyy-mm-dd hh24:mi:ss
    Now we use JodaTime’s DateTimeFormat, which use format:
    yyyy-MM-dd HH:mm:ss
    which is far more common. Going forward please use JodaTime’s DateTimeFormat in to_date and to_timestamp functions.

Features#

  • Add explode, posexplode, inline dataframe methods

  • Add struct function

  • Support schema discovery for Iceberg and Delta

  • Add if_exists clause to Table.drop_table

Fixes#

  • Do not accept a boolean value for ‘on’ parameter in ‘join’ method

Under the Hood#

  • Compress logical plan if server supports it

Dependencies#

  • Update trino-python-client to 0.328.0

  • Update dependencies

PyStarburst 0.7.0 - February 21, 2024#

Features#

  • Add DataFrame.is_empty() method

  • Adds melt alias to unpivot

  • Add DataFrame.to() method

  • Add DataFrame.with_columns_renamed() and modified with_column_renamed accordingly

  • Add LambdaFunctionExpression

  • Add transform function

  • Add zip_with function

  • Added filter function

  • Added any_match function

  • Added all_match function

  • Added array_sort function

  • Added reduce function

  • Added transform_keys, transform_values functions

  • Added map_filter function

  • Added map_zip_with function

  • Support type coercion

  • Add json_array_length function

  • Add is_not_null function

  • Add to_json function

  • Support type coercion

  • Add from_json function

  • Add programmatic access to Starburst schema discovery feature, using discover method

  • Add get_json_object function

  • Add json_tuple function

Under the Hood#

  • Use EXECUTE IMMEDIATE for prepared statements by default

  • Allow up to 3 arguments for LambdaFunctionExpression

  • Give random names for lambda parameters

  • Add tests against Python 3.12

  • Lambda functions retain context

  • Remove source_plan from TrinoPlan

Dependencies#

  • Remove typing-extensions dependency

  • Remove backports-zoneinfo dependency

  • Remove testcontainers dependency

  • Update dependencies

  • Update trino to 0.327.0

PyStarburst 0.6.3 - November 09, 2023#

Under the Hood#

  • Conda release

PyStarburst 0.6.2 - October 18, 2023#

Features#

  • Add support for stack function

  • Implement to_pandas() method

  • Add map functions: create_map, map_from_arrays, map_keys, map_values, map_entries, map_from_entries, map_concat

  • Add array functions: reverse, flatten, sort_array, shuffle, sequence, array_prepend, array_compact, get

  • Add support for unpivot function

Fixes#

  • Fix to_timestamp to accept not only VARCHAR as an argument

Under the Hood#

  • PyPi release

PyStarburst 0.5.0 - August 17, 2023#

Features#

  • Add support for regexp_split

  • Add collect_list and collect_set

  • Add support for TIME WITH TIME ZONE

  • Implement array functions: array_distinct, array_min, array_max, array_union, array_except, array_slice, size, array_repeat, array_insert, array_join, array_remove, element_at, arrays_zip

  • Support for aggregate functions: max_by, min_by, product

  • Support tuple as struct

  • Support ilike column function

  • Support math functions