WF1 Split function

Processing this data set can take time. To speed the reorganization process, you can separate the secondary index records from the logical relationship records. The logical relationship records are written to PRPURWF1. The secondary index records are written to individual files for parallel sort processing.

You can use the WF1 Split function in either of the following ways:

• Include a DFSURWF1 DD statement and an HLDURIDX DD statement in the Reload function job-step JCL. The data set that is specified in the DFSURWF1 DD statement is used for logical relationship processing. The data set specified by the HLDURIDX DD statement is used for secondary index processing. These data sets are not interchangeable.
• Specify the OUTDDX keyword in the Reload function job and the INDDX keyword in the Index Build function keyword.

Some reasons for using this function follow:

• By splitting the two sets of WF1-type records, you can process the logical relationship records and the secondary index records concurrently. Concurrent processing shortens the total load or reorganization cycle. Each index can be written to a separate file, so the indexes can be created concurrently.
• If the amount of data in the WF1 data set exceeds the DASD sort work space, you might be forced to use tape for sorting. When the one large data set is split into two smaller data sets, the sort product requires less DASD sort work space.
• The WF1 Split function generates a record count of the logical relationship and secondary index records. This count can be passed to a prefix resolution utility by using the PARM='FILSZ=n' parameter. Providing the sort product with the input file size improves sort performance.

The data sets containing the secondary index records are passed to the BUILD function. The data set containing the logical relationship records is passed to the Prefix Resolution/Update function or the IMS Prefix Resolution utility. In either case, the prefix resolution process runs significantly faster and uses less DASD sort work space.

To use the WF1 Split function, specify the following keywords:

• The Reload function OUTDDX keyword—place secondary index work file records in data sets other than DFSURWF1
• The Index Build function INDDX keyword—read secondary index records from a data set other than DFSURWF1

Use the HLDURIDX DD statement to split logical relationship and secondary index records without creating a separate data set for each index. If you want to split the secondary index records from the logical relationship records, specify a DFSURWF1 DD statement and an HLDURIDX DD statement.

The HLDURIDX data set is optional; if used, it contains all secondary index records. The DFSURWF1 data set contains logical relationship records. The HLDURIDX data set can be used as input to Index Build function if you specify the INDD keyword.