Overview

• Currently, only MRS 3.5.0-LTS and later versions support Hudi hidden partitioning.
• You can enable hidden partitioning and define its rules only during table creation in Spark or Flink. Once the table is created, hidden partitioning cannot be disabled and its rules cannot be modified.

The Hudi uses hidden partitioning to implicitly convert the original columns in a table and generate partition column values without adding columns. This allows you to flexibly create partition levels and partition directories that improve the query pruning efficiency and query performance.

Currently, Hudi hidden partitioning supports only write operations with Flink and Spark and read operations with Flink, Spark, and Hetu. Hive reads are not supported.

The following types are supported during conversation.