Release 436 (11 Jan 2024)#
General#
- Improve performance by not generating redundant predicates. (#16520) 
- Fix query failure when invoking the - json_tablefunction. (#20122)
- Fix query hang when a SQL routine dereferences a row field. (#19997). 
- Fix potential incorrect results when using the - ST_Centroid()and- ST_Buffer()functions for tiny geometries. (#20237)
Delta Lake connector#
- Add support for querying files with corrupt or incorrect statistics, which can be enabled with the - parquet_ignore_statisticscatalog session property. (#20228)
- Improve performance of queries with selective joins on partition columns. (#20261) 
- Reduce the number of requests made to AWS Glue when listing tables, schemas, or functions. (#20189) 
- Fix incorrect results when querying Parquet files containing column indexes when the query has filters on multiple columns. (#20267) 
ElasticSearch connector#
- ⚠️ Add support for ElasticSearch version 8, and remove support for ElasticSearch version 6. (#20258) 
- Add OpenSearch connector. (#11377) 
Hive connector#
- Reduce the number of requests made to AWS Glue when listing tables, schemas, or functions. (#20189) 
- Fix failure when reading certain Avro data with Union data types. (#20233) 
- Fix incorrect results when querying Parquet files containing column indexes when the query has filters on multiple columns. (#20267) 
Hudi connector#
- Add support for enforcing that a filter on a partition key must be present in the query. This can be enabled by with the - hudi.query-partition-filter-requiredconfiguration property or the- query_partition_filter_requiredcatalog session property. (#19906)
- Fix incorrect results when querying Parquet files containing column indexes when the query has filters on multiple columns. (#20267) 
Iceberg connector#
- Add support for querying files with corrupt or incorrect statistics, which can be enabled with the - parquet_ignore_statisticscatalog session property. (#20228)
- Improve performance of queries with selective joins on partition columns. (#20212) 
- Reduce the number of requests made to AWS Glue when listing tables, schemas, or functions. (#20189) 
- Fix potential loss of data when running multiple - INSERTqueries at the same time. (#20092)
- Fix incorrect results when providing a nonexistent namespace while listing namespaces. (#19980) 
- Fix predicate pushdown not running for Parquet files when columns have been renamed. (#18855) 
SQL Server connector#
- Fix incorrect results for - DATETIMEOFFSETvalues before the year 1400. (#16559)