Rate This Document
Findability
Accuracy
Completeness
Readability

Change Description

The Kunpeng BoostKit for Big Data OmniOperator uses one base to support different engines (such as Spark and openLooKeng), reducing repeated optimization work, fully exploring common and heterogeneous computing power, and promoting the Kunpeng ecosystem.

New Features

  • Supports UDFs.
  • Implemented vectorized instruction acceleration in OmniOperator.
  • Enabled decimal overflow and cast conversion for expressions in OmniOperator.
  • Enhanced basic Spark functions and the aggregation and Join operators in OmniOperator.
  • Adapted to the Spark AQE feature.
  • Added support for the Short type in OmniVec.

Modified Features

None

Removed Features

None