Writing partitioned data

In a job that uses multiple nodes, records that arrive on the input link of the connector are distributed across multiple nodes. Then, the records are written in parallel from all of the nodes to the target database.

About this task

The default partition type is Auto, which selects the partition type based on the various settings for the stages in the job. In general, instead of using Auto, it is better to select a partition type based on your knowledge about the actual data and the target table that the connector writes to at run time. In particular, if the target table is range-partitioned or list-partitioned, select Oracle connector. When the Oracle connector partition type is selected, the connector partitions the input records so that each node writes rows to the partition that is associated with that node.

Procedure

  1. On the job design canvas, double-click the Oracle Connector stage.
  2. On the Input page, select the input link.
  3. On the Partitioning page, select a partition type.