In most cases, a single thread is used to retrieve data from JDBC data sources, such as AnalyticDB for MySQL, ApsaraDB RDS for SQL Server, ApsaraDB RDS for PostgreSQL, AnalyticDB for PostgreSQL, and Druid. To improve read performance, you can use one of the following methods:

  • If your RDS performance is excellent, you can use jdbc-splits-scan to specify multiple threads to concurrently retrieve data.
/*+jdbc-scan-splits=5*/ select * from tbl1;
  • By default, the execution engine selects a random column to split columns. You can use the jdbc-split-column parameter to specify the column that the execution engine selects. Note that the column you specify must be of INT or BIGINT, or of the CHAR or VARCHAR type with indexes. If you specify a column of other types, the execution engine automatically skips this parameter.
/*+jdbc-split-column=columnName1*/ select * from tbl1;