If the output dataset can be a KSDS, how about defining the KSDS then
repro the inputs to the KSDS?

On Fri, Sep 2, 2016 at 6:19 PM, Bill Woodger <[email protected]> wrote:
> On Fri, 2 Sep 2016 15:00:28 -0500, Edward Gould <[email protected]> 
> wrote:
>
> I think I can assume this is me.
>
> I already acknowledged it *can* be done in an exit.
>
> What I'm saying is that last time I tried it, it was not a good way to do it.
>
> My exit was in COBOL. COBOL IO for a KSDS is considerably slower than DFSORT 
> itself does it, and REPRO is even better at flattening a KSDS.
>
> On top of the IO differences, you have the overhead of the exit.
>
> Just because it can be done, doesn't mean it is a good idea.
>
> OK, I could at least have recommended an evaluation, but I'd expect 
> REPRO-to-flat vs EXIT for multiple KSDS to have substantially fewer EXCPs 
> across the whole process.
>
> What I don't want people to do is to blindly jump onto the idea that an exit 
> is a great way to do this.
>
>>I think the Sort teams response answers it.
>>I think Bill Wood’s response was off the mark.
>> E15 allows the exit to read multiple types of DS’s (assuming all things 
>> being equal i.e. LRECL for example) .
>>I have never experimetted with SORT reading a VSAM ds so I will pass on 
>>commenting.
>>Since sort can do both let the sort do it.
>>
>>Ed
>>----------------------------------------------------------------------
>>For IBM-MAIN subscribe / signoff / archive access instructions,
>>send email to [email protected] with the message: INFO IBM-MAIN
>
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN



-- 
Mike A Schwab, Springfield IL USA
Where do Forest Rangers go to get away from it all?

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to