$READGRP
                  Genericize Data Set Filter List             08/05/04

If you are using dsname filter lists produced by OFFSITE or HSCTMC, it
is important to the accuracy of your study that you "genericize" the
dsnames in the filter list before using them in any analysis runs.

If your list of dsnames for WORKLOAD  LARGE  came from FINDLRG and
includes names like:
PROD.PAYROLL.DATA.D99361.T123527
PROD.PAYROLL.DATA.D99362.T133512
PROD.PAYROLL.DATA.D99363.T141231
PROD.PAYROLL.DATA.D99364.T093421

but your SMF data is for the previous month and has names for different
actual days:
PROD.PAYROLL.DATA.D99341.T093527
PROD.PAYROLL.DATA.D99342.T073527
PROD.PAYROLL.DATA.D99343.T103527

then none of the SMF data will be put into LARGE since none of the
specific names match.

What you should do is "genericize" the filter list to:
PROD.PAYROLL.DATA.D99*.T*  before doing the analysis so all SMF data
sets matching the pattern will be put into LARGE.

Before now, genericizing these lists could be very time consuming
especially when they contain thousands of entries.  But, we now have a
tool that will take your input list and apply some "genericizing" rules
that you define and write out a new list a fter it deletes any
duplicates that may have been created.  Review the new list and see if
there are any additional GRPDSN parameters that you need to define, and
simply re-run the job until you are satisfied with the new list.

The JCL member has an example of how to create the GRPDSN parameters,
like: GRPDSN=PROD.".D99*.T*


 1.  GRPDSN   - Assembler code that uses your rules to genericize a
                list of dsnames.

     Creates:   USERID.SITENAME.GRPDSN.L&runid