Parallel execution and partitioned data for the ODBC connector

For parallel jobs, the connector supports parallel reads.

You set Enable partitioning to Yes to allow parallel execution. At run time, when you specify that the data is not partitioned, the connector uses an unmodified WHERE clause to return all of the rows of data. When you specify that the data is partitioned, the connector modifies the WHERE clause to return a subset of the total number of rows. The connector performs the WHERE clause modification on each query that is on each processing node. Each node then receives a unique subset of rows. The subsets of rows are determined by the partitioning method that you specify in the Partitioning method property.