Release 451 (27 Jun 2024)#

General#

  • Add support for configuring a proxy for the S3 native filesystem with the s3.http-proxy.username, s3.http-proxy.password, s3.http-proxy.non-proxy-hosts, and s3.http-proxy.preemptive-basic-auth configuration properties. (#22207)

  • Add support for the t_pdf() and t_cdf() functions. (#22507)

  • Improve performance of reading JSON array data. (#22379)

  • Improve performance of certain queries involving the row_number(), rank(), or dense_rank() window functions with partitioning and filters. (#22509)

  • Fix error when reading empty files with the native S3 file system. (#22469)

  • Fix rare error where query execution could hang when fault-tolerant execution is enabled. (#22472)

  • Fix incorrect results for CASE expressions of the form CASE WHEN ... THEN true ELSE false END. (#22530)

Delta Lake connector#

  • Improve performance of reading from Parquet files with large schemas. (#22451)

Hive connector#

  • Improve performance of reading from Parquet files with large schemas. (#22451)

Hudi connector#

  • Improve performance of reading from Parquet files with large schemas. (#22451)

Iceberg connector#

  • Add support for incremental refresh for basic materialized views. (#20959)

  • Add support for adding and dropping fields inside an array. (#22232)

  • Add support for specifying a resource prefix in the Iceberg REST catalog. (#22441)

  • Add support for partitioning on nested ROW fields. (#15712)

  • Add support for writing Parquet Bloom filters. (#21570)

  • Add support for uppercase characters in the partitioning table property. (#12668)

  • Improve performance of reading from Parquet files with large schemas. (#22451)

Kudu connector#

  • Add support for the Kudu DATE type. (#22497)

  • Fix query failure when a filter is applied on a varbinary column. (#22496)

SPI#

  • Add a Connector.getInitialMemoryRequirement() API for pre-allocating memory during catalog initialization. (#22197)