Release 2.1.9

Dear Community, Apache Doris version 2.1.9 is now available, featuring improved SQLHash calculation, enhanced query results accuracy, and new metrics for better storage management. This update also resolves critical bugs across multiple areas for a more robust data management experience.

• Quick Download

• GitHub Release

Behavior Changes

• The SQLHash in Audit Log is now accurately calculated per SQL query, resolving the issue of identical hashes in a single request. #48242
• Query results match ColumnLabelName exactly. #47093
• User property variables now take precedence over session variables. #47185

New Features

Storage Management

• Disallow renaming partition columns. #47596

Others

• Added FE monitoring metrics for Catalogs, Databases, and Tables counts. #47891

Improvements

Inverted Index

• Support for ARRAY type in VARIANT inverted indexes. #47688
• Profile now shows performance metrics for each filter condition. #47504

Query Optimizer

• Support for using SELECT * in aggregate queries with only aggregation key columns. #48006

Storage Management

• Enhanced CCR for binlog recycling and small file transfer efficiency, and robustness in chaotic environments. #47547 #47313 #45061
• Enhanced import error messages to be more specific. #47918 #47470

Bug Fixes

Lakehouse

• Fixed BE krb5.conf path configuration issue. #47679
• Prevented SELECT OUTFILE statement retries to avoid duplicate data export. #48095
• Fixed JAVA API access to Paimon tables. #47192
• Resolved writing to Hive tables with s3a:// storage location. #47162
• Fixed the issue of Catalog's Comment field not being persisted. #46946
• Addressed JDBC BE class loading leaks under certain conditions. #46912
• Resolved high version ClickHouse JDBC Driver compatibility with JDBC Catalog. #46026
• Fixed BE crash when reading Iceberg Position Delete. #47977
• Corrected reading MaxCompute table data under multi-partition columns. #48325
• Fixed reading Parquet complex column types errors. #47734

Inverted Index

• Fixed ARRAY type inverted index null value handling. #48231
• Resolved BUILD INDEX exception for newly added columns. #48389
• Corrected UTF8 encoding index truncation issues leading to errors. #48657

Semi-structured Data Types

• Fixed array_agg function crashes under special conditions. #46927
• Resolved JSON import crashes due to incorrect chunk parameters. #48196

Query Optimizer

• Fixed constant folding issues with nested time functions like current_date. #47288
• Addressed non-deterministic function result errors. #48321
• Resolved CREATE TABLE LIKE execution issues with on update column properties. #48007
• Fixed unexpected planning errors for materialized views of aggregate model tables. #47658
• Resolved PreparedStatement exceptions due to internal ID overflow. #47950

Query Execution Engine

• Resolved query hang or null pointer issues when querying system tables. #48370
• Added DOUBLE type support for LEAD/LAG functions. #47940
• Fixed query errors when case when conditions exceed 256. #47179
• Corrected str_to_date function errors with spaces. #48920
• Fixed split_part function errors during constant folding with ||. #48910
• Corrected log function result errors. #47228
• Resolved core dump issues with array / map functions in lambda expressions. #49140

Storage Management

• Fixed memory corruption issues during import of aggregate tables. #47523
• Resolved occasional core dump during MoW import under memory pressure. #47715
• Fixed potential duplicate key issues with MoW during BE restart and schema change. #48056 #48775
• Corrected group commit and global column update issues with memtable promotion. #48120 #47968

Permission Management

• No longer throws PartialResultException when using LDAP. #47858