3.0.2-palantir.2

• [SPARK-34803][PYSPARK] Pass the raised ImportError if pandas or pyarrow fail to import (#745)
• Support for reading multiple sorted files per bucket (#742)

3.0.0-palantir.94

• [SPARK-26713][CORE][2.4] Interrupt pipe IO threads in PipedRDD when task is finished (#741)

3.0.2-palantir.1

• [BUILD] Set hadoop-palantir profile for publish functions (#739)

3.0.2-palantir.0

This is palantir/spark’s first “full” Spark 3 release. It contains all of Spark 3 and has Palantir’s changes rebased on top of the upstream v3.0.2 tag.

• For a comprehensive list of changes in Spark 3 consult the Migration Guide (Apache).
• For the updated list of differences between our fork and upstream see our FORK.md
• You can find the full set of palantir/spark changes compared against tag v3.0.2 in #737

3.0.0-palantir.93

• Use non-batch scans during sorted bucketed reads (#738)

3.0.0-palantir.0.alpha.6

Do not use.

3.0.0-palantir.0.alpha.5

Share secret name between drivers and executors when mounting files u…

3.0.0-palantir.92

• [SPARK-30371][K8S] Add spark.kubernetes.driver.master conf (#736)

3.0.0-palantir.91

• Support for reading multiple sorted files per bucket (#731)
• [SPARK-32753][SQL] Only copy tags to node with no tags (#732)
• [SPARK-27134][SQL] array_distinct function does not work correctly with columns containing array of array (#734)

3.0.0-palantir.0.alpha.4

Bump py4j to 0.10.9.1.