Release 481 (11 May 2026)#

General#

  • Add support for casting boolean to number. (#28879)

  • Add support for the number type in Python UDFs. (#28921)

  • Add support for casting between number and json. (#28394)

  • Improve performance of json_value and json_table by evaluating ON EMPTY and ON ERROR clauses lazily. (#28969)

  • Add support for DESCRIBE OUTPUT with inline queries, allowing direct description of query results without requiring a PREPARE statement. For example, DESCRIBE OUTPUT (SELECT * FROM nation). (#28002)

  • Add support for the NEAREST clause for approximate matches in joins. (#21759)

  • Add support for binding parameters in WITH SESSION, SET SESSION, and CALL statements. (#29053)

  • Add support for persisting external authentication tokens to disk (in ~/.trino/), allowing reuse across separate JDBC client processes. Set externalAuthenticationTokenCache=SYSTEM to enable. (#28783)

  • Include connector split source metrics in io.trino.spi.eventlistener.QueryInputMetadata#connectorMetrics. (#28870)

  • Replace the Esri geometry library with JTS to improve interoperability with the broader spatial ecosystem. (#27881)

  • ⚠️ Breaking change: Reject WKT input that does not conform to OGC standards. Some inputs that were silently accepted before will now fail. (#27881)

  • ⚠️ Breaking change: Change ST_Union() to return an empty geometry collection instead of NULL for empty inputs. (#27881)

  • ⚠️ Breaking change: Stop inserting vertices at intersection points for point-on-line unions in ST_Union(). (#27881)

  • Improve compatibility of file system exchange with Azure containers that have hierarchical namespaces enabled. (#29042)

  • Improve performance of queries with simple AND and OR predicates that have highly selective terms. (#24336)

  • Reduce excessive memory usage caused by suboptimal join ordering for queries on columns with unknown statistics. (#29157)

  • Fix query failures and transaction errors caused by race conditions when dynamic catalogs are dropped concurrently with system.jdbc or system.metadata queries. (#28017)

  • Fix incorrect results when using json_parse() or JSON type literals on documents containing numbers with more than 16 significant digits in the decimal portion. (#28867)

  • Fix failure when executing table procedures on tables with uppercase column names. (#28970)

  • Fix failure when executing ALTER TABLE EXECUTE OPTIMIZE with OR predicates on partitioned timestamp with time zone columns. (#27136)

  • Fix incorrect freshness check for materialized views whose definition contains non-deterministic functions. (#28682)

  • Fix failure when executing DESCRIBE OUTPUT with [VERSION | TIMESTAMP] AS OF clauses. (#29077)

JDBC driver#

  • Add support for the variant type. (#29046)

  • Add support for transparent OAuth2 token refresh when using the JDBC accessToken connection property against a server configured with http-server.authentication.oauth2.refresh-tokens=true. Previously, such connections failed with 401 Unauthorized once the embedded access token expired. (#29264)

  • Fix failure when calling setBigDecimal() with a BigDecimal value whose toString() representation uses scientific notation, such as 0E-10. (#23523)

CLI#

  • Add support for the variant type. (#29046)

ClickHouse connector#

  • Add support for reading all ClickHouse DECIMAL columns. (#28873)

Delta Lake connector#

  • ⚠️ Breaking change: Remove legacy object storage support for Azure Storage, Google Cloud Storage, IBM Cloud Object Storage, S3, and S3-compatible systems. Use native file system support for object storage. fs.hadoop.enabled applies only to HDFS. See legacy file system support for migration details. (#24878)

  • Add support for configuring the Azure connection-pool idle time and HTTP request timeout via the azure.connection-pool-max-idle-time and azure.http-request-timeout configuration properties. (#29284)

  • Reduce memory usage when writing files to S3. (#28488)

  • Reduce query planning time by reading metadata and protocol information from Delta checksum files. This optimization can be disabled by setting the delta.load-metadata-from-checksum-file configuration property or load_metadata_from_checksum_file session property to false. (#28381)

  • Fix DELETE deleting incorrect rows on Delta Lake tables backed by Parquet files with column indexes, particularly when the table was written by Apache Spark. (#28885)

Hive connector#

  • ⚠️ Breaking change: Remove legacy object storage support for Azure Storage, Google Cloud Storage, IBM Cloud Object Storage, S3, and S3-compatible systems. Use native file system support for object storage. fs.hadoop.enabled applies only to HDFS. See legacy file system support for migration details. (#24878)

  • Add support for Esri GeoJSON. (#28859)

  • Add support for configuring the Azure connection-pool idle time and HTTP request timeout via the azure.connection-pool-max-idle-time and azure.http-request-timeout configuration properties. (#29284)

  • Improve performance of queries with date partition predicates on Hive tables in the Glue catalog. (#28817)

  • Reduce memory usage when writing files to S3. (#28488)

  • Fix failure when creating tables in the Hive metastore version 3.1. (#28798)

  • Fix failure when parsing bucket numbers on tables backed by files without a bucket name. (#28632)

Hudi connector#

  • Add support for configuring the Azure connection-pool idle time and HTTP request timeout via the azure.connection-pool-max-idle-time and azure.http-request-timeout configuration properties. (#29284)

Iceberg connector#

  • ⚠️ Breaking change: Remove legacy object storage support for Azure Storage, Google Cloud Storage, IBM Cloud Object Storage, S3, and S3-compatible systems. Use native file system support for object storage. fs.hadoop.enabled applies only to HDFS. See legacy file system support for migration details. (#24878)

  • Add experimental support for the variant type for Iceberg v3 tables. Older CLI versions will render the value as JSON unless updated. (#24538)

  • Add support for reading and writing timestamp(9) and timestamp(9) with time zone in Iceberg v3 tables. (#27835)

  • Add support for Azure vended credentials in Iceberg REST catalog. (#23238)

  • Add support for specifying HTTP headers in REST catalog. (#24236)

  • Add support for execution metrics while running the add_files and add_files_from_table procedures. (#28996)

  • Add support for execution metrics while running the optimize procedure. (#28992)

  • Add more columns to $files system table. (#29044)

  • Add added_snapshot_id column to $files system table. (#28911)

  • Add support for configuring the Azure connection-pool idle time and HTTP request timeout via the azure.connection-pool-max-idle-time and azure.http-request-timeout configuration properties. (#29284)

  • Add support for refreshable vended credentials for the Iceberg REST catalog (S3, GCS, Azure). (#28998)

  • Reduce memory usage when writing files to S3. (#28488)

  • Fix failure when reading the $entries or $all_entries metadata tables. (#29211)

Lakehouse connector#

  • ⚠️ Breaking change: Remove legacy object storage support for Azure Storage, Google Cloud Storage, IBM Cloud Object Storage, S3, and S3-compatible systems. Use native file system support for object storage. fs.hadoop.enabled applies only to HDFS. See legacy file system support for migration details. (#24878)

  • Add support for configuring the Azure connection-pool idle time and HTTP request timeout via the azure.connection-pool-max-idle-time and azure.http-request-timeout configuration properties. (#29284)

  • Reduce memory usage when writing files to S3. (#28488)

MongoDB connector#

  • Fix incorrect results when reading JSON columns containing numbers with more than 16 significant digits in the decimal portion. (#28867)

MySQL connector#

  • Fix incorrect results when reading JSON columns containing numbers with more than 16 significant digits in the decimal portion. (#28867)

Pinot connector#

  • Fix incorrect results when reading JSON columns containing numbers with more than 16 significant digits in the decimal portion. (#28867)

PostgreSQL connector#

  • Improve performance of queries involving COALESCE by pushing expression computation to the underlying database. (#11535)

  • Fix incorrect results when reading JSON columns containing numbers with more than 16 significant digits in the decimal portion. (#28867)

SingleStore connector#

  • Fix incorrect results when reading JSON columns containing numbers with more than 16 significant digits in the decimal portion. (#28867)

SQL Server connector#

  • Add support for the json type. (#28082)

  • Fix permission denied failure when reading tables. (#29244)

SPI#

  • Add variant type defined by the Iceberg specification. (#24538)

  • Add a tableBranch parameter to the ConnectorAccessControl.checkCanXxx table-level methods. (#29179)

  • Add support for pushing down COALESCE expressions in connectors. (#28984)

  • ⚠️ Breaking change: Remove the deprecated getObject and appendTo methods from the Type interface. (#29003)