Mainframe, MVS and zOS Discussion

Hi,

What is the use of PARM='CENTWIN='? In one of the jobs we're using PARM='CENTWIN=1970'.

When I search for it, I came to know that it is used to provide a century window between years but could not understand for 1970 what it would be?

The SORT card is SORT FIELDS=COPY and the this sort-step has 35 files concatenated to each other to create a single output file. This conuses me what benefit/drawbacks CENTWIN can provide?

Kindly advise.

it is a way of interpreting and sorting correctly dates expressed using a 2 digit year

a 1970 window specification would mean that

for
YY > 70 the 4 digit year would be interpreted as 19YY
for
YY <= 70 the 4 digit year would be interpreted as 20YY

read the manuals for the details for YY = 70 ( could the other way around )

Thanks, you are correct Enrico. I read more about CENTWIN and came to know that CENTWIN is an alias for DFSORT's Y2PAST option (and applicable in SyncSort too) which is used to set the "century window" you want to use with DFSORT's Y2x formats.

I'm thinking if it'll have something to with CPU utilization or REGION overhead if used unncessarily. I've many SORT jobs here which are using it and the SORT job is used only for copy operation and no date calculations are involved.

I've executed a job with 200K+ records using PARM='CENTWIN=1970' and witout it. There is a difference of .01Sec. Now I'm not sure if it's worth to invest time in this direction...

the overhead will be negligible ...
it is not worth to waste time worrying about the overhead when there might be the risk of a wrong sort sequence!

I'd not expect it to make a difference if there is no date-processing anyway.

Concatenate your file five times on SORTIN (so you get 1m+ records without taking up space) and compare.

0.01 difference is significant when original CPU usage is 0.02. Without knowing the total usage, it is not possible for us to judge on a difference of 0.01.

Thanks for the directions enrico and William. I'll work on these and let you know.