Collecting distribution statistics for column groups

When RUNSTATS collects distribution statistics for columns groups, the utility invokes a sort program to sort the distribution statistics. This sort requires its own work data set. The DD name is STATWK01.

About this task

You can let this data set be dynamically allocated through the sort program, or you can allocate the data set through a DD statement in the job JCL.

If you need to control the size or placement of the data sets, use the JCL statements to allocate STATWK01.

If the column group contains one column, you can potentially improve RUNSTATS performance by specifying the STATCLGMEMSRT option or changing the value of the STATCLGSRT subsystem parameter. You can use these options to avoid the external data sort that is used for column group processing.

Procedure

To collect distribution statistics for column groups:

To let the work data set be dynamically allocated, remove the STATWK01 DD statements from the job and allocate the UTPRINT statement to SYSOUT.
To let the sort program dynamically allocate this data set, specify the SORTDEV option in the RUNSTATS utility control statement.