On Mon, 25 Mar 2013 12:28:00 -0400, Thomas David Rivers wrote:

>Bill Godfrey wrote:
>
>>On Mon, 25 Mar 2013 09:46:56 -0400, Thomas David Rivers wrote:
>>
>>
>>
>>>I'm reworking some code to avoid multiple OPENs for
>>>reading PDS members..
>>>
>>>But, I will still be just reading one member at a time,
>>>and I don't know the member names a-priori.
>>>
>>>The DFSMS documentation states that using BLDL + FIND
>>>(and using the ttr,C form for FIND) results in a performance
>>>improvement... over simply pointing FIND at the DCB.
>>>
>>>
>>
>>I am curious to know where the documentation says that. Can you be more 
>>specific?
>>
>>
>>
>Hi Bill,
>
> I'm looking at z/OS V1R12.0 DFSMS Using Data Sets (SC26-7410-10)
>under the heading "BLDL - Construct a Directory Entry List".  Paragraph
>4 says:
>
>    You can improve retrieval time by directing a subsequent FIND
>    macro to the BLDL list rather than to the directory to locate
>    the member to be processed.
>
> I can readily believe that statement if the BLDL list has more than
> one entry... hence my question...
>
Thanks. That was my understanding of how the improved retrieval time would be 
documented. Your re-wording made me wonder if I had missed something. The 
sentence is referring to the difference between the 2 types of FIND after a 
BLDL and does not make any comparison to a FIND without a BLDL. Obviously if a 
BLDL list is created with multiple members, an improvement can be realized, but 
that's not what this sentence is about. I would expect that a FIND without a 
BLDL is just negligibly faster than BLDL + FIND(C) if the BLDL list is only 
going to have one member name, since a BLDL list is not needed.

Bill

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

Reply via email to