This repository has been archived by the owner on Mar 11, 2024. It is now read-only.
Releases: danielgafni/dagster-polars
Releases · danielgafni/dagster-polars
Release v0.3.1
What's Changed
- 🐛 fix sometimes wrong type check when writing df by @danielgafni in #55
Full Changelog: v0.3.0...v0.3.1
Release v0.3.1a0
Full Changelog: v0.3.0...v0.3.1a0
Release v0.3.0
✨ What's Changed
- feat: add lazyframe support for parquet and delta by @ion-elgreco in #52.
sink_df_to_path
method can be used to implementLazyFrame
sinking cloud_storage_options
config field added. Supports passing values withdagster.EnvVar
, for example:
PolarsDeltaIOManager(cloud_storage_options={"aws_secret_access_key": EnvVar("AWS_SECRET_ACCESS_KEY")})
- removed legacy pyarrow parquet reader
💥 Breaking Changes
- minimal compatible
polars
version has been increased to0.20.0
BasePolarsUPathIOManager
- renameddump_df_to_path
towrite_df_to_path
🤝 New Contributors
- @ion-elgreco made their first contribution in #52. Thank you for your work!
Full Changelog: v0.2.2...v0.3.0
Release v0.3.0a0
Testing release for LazyFrame sink support
Full Changelog: v0.2.2...v0.3.0a0
Release v0.2.2
🐛 Bug Fixes
- Due to an issue with
UPathIOManager
,PolarsDeltaIOManager
did not apply correct partition filtering when loading a single mapped partition of an asset with native DeltaLake partitioning.
A workaround for this issue has been included. - Also, it's now possible to load all asset partitions at once with native DeltaLake partitioning, by specifying a type annotation like
pl.DataFrame
orpl.LazyFrame
. No partitions filtering will be done in this case.
What's Changed
- 🐛 DeltaLake - fix loading single partitions, implement loading all partitions at once by @danielgafni in #51
Full Changelog: v0.2.1...v0.2.2
Release v0.2.1
What's Changed
- 🐛 properly overwrite DeltaLake partitions by @danielgafni in #50
Full Changelog: v0.2.0...v0.2.1
Release v0.2.0
💥 Breaking Changes
- Rename
BigQueryIOManager
->PolarsBigqueryIOManager
What's Changed
- ⭐ skip BigQueryIOManager upload if output is None or empty by @danielgafni in #46
- 📝 💥 prepare for merge into Dagster project by @danielgafni in #48
- ✨ automatically filter partitions in the IOManagers by @danielgafni in #49
Full Changelog: v0.1.5...v0.2.0
Release v0.1.5
What's Changed
- 👷 add required qa_success CI step by @danielgafni in #42
- ✨ support storing arbitrary metadata in PolarsDeltaIOManager by @danielgafni in #44
- 👷 fix matrix packages installation in CI by @danielgafni in #45
Full Changelog: v0.1.4...v0.1.5
Release v0.1.4
What's Changed
The new native polars Parquet reader is going to be used with polars >= 0.19.4
.
The old, pyarrow
reader can be enabled in PolarsParquetIOManager
config.
- 👷 add manual CI trigger by @danielgafni in #39
- 🐛 fix missing
pyarrow.dataset
import forpyarrow 13.0.0
by @seandavi in #38 - ⚡ use native parquet reader by @danielgafni in #41
New Contributors
Full Changelog: v0.1.3...v0.1.4
Release v0.1.3
What's Changed
- 👷 remove MacOS tests in CI/CD by @danielgafni in #34
- 🔊 log Delta Lake table version when loading it & make sure time-traveling works by @danielgafni in #36
Full Changelog: v0.1.2...v0.1.3