Skip to content

Releases: CrunchyData/pg_parquet

v0.2.0

17 Dec 14:34
fd51bed
Compare
Choose a tag to compare

This is pg_parquet v0.2.0, which supports Postgres(14-15) and introduces some useful features which improve overall usability. We also have few changing behavior.

Important Changes

  • pg14 and pg15 is supported (#57),
  • Improvements and fixes on copying numeric type (#65),
  • Support coercions during COPY FROM parquet. This lets it work seamlessly even if the parquet schema and table's schema is slightly different but coercible (#76),
  • Adds match_by option for COPY FROM. Parquet schema and table schema can now be matched by position or name (#80),
  • AWS_ENDPOINT_URL and AWS_SESSION_TOKEN works for s3 object store (#84).

Breaking Changes

  • Default behavior change in matching parquet and table schemas. By default, match by field position instead of field name (#80),
  • COPY TO/FROM will not work if pg_parquet extension is not created or when it is dropped (#85),
  • numeric column, which does not specify precision and scale, is written as numeric(38,9) to parquet file (#86).

Full Changelog: v0.1.0...v0.2.0