Parallel data loader support to use with the CPYFRMIMPF command

 

The Copy From Import File (CPYFRMIMPF) command supports copying the data in parallel from an import file to a to-file using multiple jobs during the copy operation. This allows you to copy data files from other platforms into a to-file quickly and easily. This is especially useful for those who use data warehousing. To use multiple jobs, your system must have the i5/OS^® operating system option DB2^® Symmetric Multiprocessing (SMP).

The number of jobs you use during the copy is determined by the DEGREE(*NBRTASKS) parameter of the Change Query Attributes (CHGQRYA) command. If the from-file has less than 50,000 records, only one job will be used regardless of the *NBRTASKS value.

The CPYFRMIMPF command (with the parallel data loader support) essentially breaks the import file into smaller portions or blocks. Each of these smaller portions is submitted in parallel, so the entire file processes at the same time. (This eliminates the latency of sequential processing.)

To maintain the same relative record numbers of the from-file in the to-file, use only one job for the copy. Specify DEGREE(*NONE).

Records are not copied in parallel when the from stream file (FROMSTMF) parameter is specified.

 

Parent topic:

Using the Copy From Import File (CPYFRMIMPF) command to copy between different systems

 

Related concepts

Using the Copy From Import File (CPYFRMIMPF) command to copy between different systems

 

Related tasks

Importing data to the System i platform when the from-file is a database file or DDM file