Changes in MySQL HeatWave 9.7.0 (2026-04-21)

Note

These release notes were created with the assistance of MySQL HeatWave GenAI.

• MySQL HeatWave GenAI

• MySQL HeatWave Lakehouse

• MySQL HeatWave

MySQL HeatWave GenAI

• The NL_SQL routine now supports Gemini and OpenAI models. (Bug #38824151)

MySQL HeatWave Lakehouse

• MySQL HeatWave Lakehouse now enables flexible data ingestion from AVRO files with enhanced support for column re-mapping with the match_columns_by parameter. You can load columns into Lakehouse tables by choosing from three column mapping modes: by order, name (case-sensitive), or name (case-insensitive). MySQL HeatWave Lakehouse now also supports selective loading of AVRO files with improved schema inference capabilities. The allow_missing_columns parameter lets you disable or enable the ability to include all unique columns by name found during inference. When you enable allow_missing_columns, MySQL HeatWave Lakehouse loads all unique columns even if they are not present in every file, and loads data with missing or extra columns without errors. These enhancements provide more control over data loading and ensure seamless integration with both auto schema inference and Guided Load processes, making it easier to load AVRO files with varying column structures.

  For more information, see Column Matching Options and Allow Missing Columns Option. (WL #17195, WL #17196)

MySQL HeatWave

• MySQL HeatWave now supports bulk loading of tables with foreign keys, ensuring that all foreign key constraints are satisfied for each record. This update enables you to successfully load data into tables with foreign key constraints, while also providing error reporting for violated constraints. Although this added functionality may impact performance due to the extra checks required, it provides a necessary trade-off for supporting tables with foreign keys, making it easier to manage complex data relationships within MySQL HeatWave.

  For more information, see Bulk Ingest Data. (WL #17159)

• MySQL HeatWave now supports the tracking of causes for loaded tables becoming stale. With this enhancement, you can easily identify and diagnose issues, and in some cases take corrective action.

  For more information, see The rpd_tables Table. (WL #17002)