The Data Integration Service can create partitions for a flat file or a file in Hadoop Distributed File System (HDFS).

You can configure a flat file data object to have either a file or command output type. When a flat file data object has the file output type, the Data Integration Service writes the target data to a flat file. If multiple threads write to the flat file target, each thread writes the target output to a separate file. Each thread uses the following format to name the file:

<output_file_name><partition_number>.out

For example, three threads might write to files named MyOutput1.out, MyOutput2.out, and MyOutput3.out.

You can configure multiple output file directories to optimize performance, or you can configure the flat file data object to write to a single merge file.

When a flat file data object has the command output type, the Data Integration Services outputs the target data to a command or to a merge command instead of a flat file or a merge file. If multiple partitions write to the flat file target, you can configure a command to process target data for a single partition or to process merge data for all target partitions.