We are running DB2 V9 in a mainframe (Z/OS) environment.

Lets say we have the following 4 column table laid out in this COLNO order:

COLNO NAME COLTYPE LENGTH
====== ==== ======= ======
01 UID INTEGER
02 EMPLNO CHAR 8
03 SSN CHAR 9
04 UPDT_TS TIMESTAMP

I have left the question of indices out for the time being; albeit, I recognize that the answer to this question may be based on the separate issue of what the indices look like.

My question is: without having to introduce a 5th column, is it possible to create a partitioning strategy for the above table that is based on the last 2 digits of the SSN column ?
Or does one have to have an additional CHAR(2) column (or something similar) that is maintained and which houses the last 2 digits of the SSN?
I ask the question independant of whether it is a pre-Version-8 "index-controlled" partition, or a Verson-8 and beyond "table-controlled" partition; although I recognize that the type of partition will probably be included in the answer. thank-you in advance to anyone who can provide an answer.

regards
db2dcs

db2dcs, you can't partition on a partial character column value (unless that partial value is at the beginning).

Thank-you StealthDBA for your answer. T'was what I thought; but I am always hoping that some day a new DB2 release will enable that capability to exist.