Help Center/ MapReduce Service/ Component Operation Guide (LTS)/ Using Hudi/ Hudi SQL Syntax Reference/ Restrictions on Using Hudi SQL
Updated on 2026-06-27 GMT+08:00
View PDF
Share

Restrictions on Using Hudi SQL

Hudi supports Spark SQL for both Data Definition Language (DDL) and Data Manipulation Language (DML) operations. This compatibility simplifies data access and management for all users, including non-technical stakeholders and analysts. This integration enables users to perform data management and operational tasks using familiar SQL syntax, removing the need to understand complex underlying technical details.

Notes and Constraints

  • You can use Spark SQL to operate Hudi on the Hudi client.
  • You can use Spark SQL to operate Hudi in JDBCServer of Spark2x.
  • The Spark2x client does not support Spark SQL operations. The Spark3.1.1 or later client support Spark SQL operations.
  • You cannot write data to Hudi tables or modify the Hudi table structure in Hive and Hetu engines. Only read operations are supported.
  • The default value of KeyGenerator in SQL is org.apache.hudi.keygen.ComplexKeyGenerator. Therefore, you need to set the KeyGenerator value to that of SQL when data is written in DataSource mode.
  • Only primary MOR tables can be modified. MOR tables suffixed with ro or rt are used only for querying.
  • In an SQL statement, a column alias must be different from the original field name.