Changes in HeatWave 8.3.0-u2 (2024-02-26, General Availability)

• HeatWave AutoML

• HeatWave Lakehouse

• Functionality Added or Changed

HeatWave AutoML

• HeatWave AutoML adds prediction intervals to the forecasting task. (WL #16072)

• HeatWave AutoML adds the PCA and GLOF algorithms to the anomaly detection task. (WL #16104)

HeatWave Lakehouse

• HeatWave Lakehouse now supports the JSON data type. (WL #16037)

• HeatWave Lakehouse now supports Guided Load to perform pre-load validation checks and schema inference. (WL #16004)

• HeatWave Lakehouse now supports the following:

  • Point-in-time-recovery.

  • High Availability.

  • Read Replication.

  • Outbound Replication.

  (WL #15926)

• HeatWave Lakehouse now supports Auto Data Compression. This dynamically determines which compression algorithm to use for each column based on its data characteristics. This provides the optimum balance between memory usage and query performance. See: Data Load. (WL #16061)

Functionality Added or Changed

• HeatWave primary and secondary engines now support the HyperLogLog, HLL() aggregate function. This is similar to COUNT(DISTINCT), but faster and with user defined precision. This is only available in HeatWave, see: Aggregate Functions. (WL #15992)

• HeatWave can now reload all tables, and unload all tables. (WL #16023)

• HeatWave secondary engine now supports the TABLESAMPLE clause to use with SELECT statements. (WL #15902)