Release 434 (29 Nov 2023)

General

• Add support for a FILTER clause to the LISTAGG function. ({issue}19869)
• {{breaking}} Rename the query.max-writer-tasks-count configuration property and the related max_writer_tasks_count session property to query.max-writer-task-count and max_writer_task_count. ({issue}19793)
• Improve performance of INSERT ... SELECT queries that contain a redundant ORDER BY clause. ({issue}19916)
• Fix incorrect results for queries involving comparisons between double and real zero and negative zero. ({issue}19828)
• Fix performance regression caused by suboptimal scalar subqueries planning. ({issue}19922)
• Fix failure when queries on data stored on HDFS involve table functions. ({issue}19849)
• Prevent sudden increases in memory consumption in some queries with joins involving UNNEST. ({issue}19762)

BigQuery connector

• Add support for reading json columns. ({issue}19790)
• Add support for DELETE statement. ({issue}6870)
• Improve performance when writing rows. ({issue}18897)

ClickHouse connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

Delta Lake connector

• {{breaking}} Disallow invalid configuration options. Previously, they were silently ignored. ({issue}19735)
• Improve performance when reading large checkpoint files on partitioned tables. ({issue}19588, {issue}19848)
• Push down filters involving columns of type timestamp(p) with time zone. ({issue}18664)
• Fix query failure when reading Parquet column index for timestamp columns. ({issue}16801)

Druid connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

Hive connector

• Add support for columns that changed from timestamp to date type. ({issue}19513)
• Fix query failure when reading Parquet column index for timestamp columns. ({issue}16801)

Hudi connector

• Fix query failure when reading Parquet column index for timestamp columns. ({issue}16801)

Iceberg connector

• {{breaking}} Remove support for legacy table statistics tracking. ({issue}19803)
• {{breaking}} Disallow invalid configuration options. Previously, they were silently ignored. ({issue}19735)
• Fix query failure when reading Parquet column index for timestamp columns. ({issue}16801)
• Don't set owner for Glue materialized views when system security is enabled. ({issue}19681)

Ignite connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

MariaDB connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

MySQL connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

Oracle connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

Phoenix connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

PostgreSQL connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)
• Prevent possible query failures when join is pushed down. ({issue}18984)

Redshift connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)
• Prevent possible query failures when join is pushed down. ({issue}18984)

SingleStore connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)

SQL Server connector

• Add support for separate metadata caching configuration for schemas, tables, and metadata. ({issue}19859)
• Prevent possible query failures when join is pushed down. ({issue}18984)

SPI

• Add bulk append methods to BlockBuilder. ({issue}19577)
• {{breaking}} Remove the VariableWidthBlockBuilder.buildEntry method. ({issue}19577)
• {{breaking}} Add required ConnectorSession parameter to the method TableFunctionProcessorProvider.getDataProcessor. ({issue}19778)