MVSFORUMS.com

bstillwa

We are attempting to use COBOL logic in CICS that was originally written to work in Batch, where the programs are now reuse=re-entrant. A table is being stored in CICS shared/main storage for use throughout the day. This table is being searched using a SEARCH ALL (logic written for Batch use).
The table is defined:
10 ECA06-L-ID-MOD-ENTRY OCCURS 3000 TIMES
ASCENDING KEY ECA06-L-FLD-ID-MOD
INDEXED BY ECA06-INDEX-L.

I have a concern that the Index which is defined within this table could be being used by more than one transaction at the same time. Is this a valid concern? Should the logic instead use a SEARCH which would use a working-storage index?

The table being searched can contain a maximum of 3000 entries, but at present actually contains only 1336 entries. For a table of this size, is there any performance gain by using SEARCH ALL over SEARCH?
_________________
Bev.

Dibakar · Posted: Mon Nov 22, 2010 12:23 pm Post subject:

bstillwa,

This is what I think, others please correct me if I am wrong.

bstillwa · Posted: Mon Nov 22, 2010 12:29 pm Post subject:

The Search Criteria are local to my programs, but the Key and Index are not local to my programs. The Key and Index are located in Shared Storage. Therefore repeating my concern: could the Key and Index be being used by more than one transaction (and/or program) at the same time?
_________________
Bev.

dbzTHEdinosauer · Posted: Mon Nov 22, 2010 12:38 pm Post subject:

bstillwa,

suggest you do some reading.

you keep saying the the key and the index are located in shared storage.
prove that statement.
_________________
Dick Brenholtz
American living in Varel, Germany

bstillwa · Posted: Mon Nov 22, 2010 12:56 pm Post subject:

Please then point me to the manual where it states that the key and index used in a binary search are never located in shared storage for a table defined and used in this manner (instead of responding with a degrading tone). Please try to be helpful. This is confusing me because the area where the table, key and index are all defined does indeed reside in shared storage. Any time this table is used, the beginiing of it is pointed into shared storage (via an address which is stored in the CWA), by use of SET ADDRESS.
_________________
Bev.

dbzTHEdinosauer · Posted: Tue Nov 23, 2010 3:46 am Post subject:

I used the tone because you have little knowledge (refuse to increase it) and only want to hear what sounds good.

the shared storage area is nothing but a data storage area.

the definitions of the data relationships (size, data type, etc...), the organization of the data, the definition of the 'keys' and the use or non-use of indexes is accompished in the COBOL program via the data defintions provided for the OCCURS clause.

That you have not grasped these fundamental concepts bodes ill for your potential to solve this problem.

you were told NO, twice, yet you persist on your course of inquiry, driven by you lack of understanding.
again, suggest you do some reading and thinking.
_________________
Dick Brenholtz
American living in Varel, Germany

papadi · Supermod Joined: 20 Oct 2009 Posts: 594 Topics: 1

computer · Posted: Mon Nov 29, 2010 5:43 am Post subject:

Hi,

Out of eagerness and not with outstanding knowledge in CICS on ADDRESS, just wanted to know what exactly is the issue?

And with what I understood, It looks like the table defined here is like a Queue which can be accessed by multiple transaction. So, if a read is done on the table with any KEY and INDEX then that value becomes local to that transaction which will not affect other transaction operation (As each transaction have got their own TWA and other space allocation).

Please correct if I going wrong anywhere.

Thanks,
Computer

Dibakar · Posted: Mon Nov 29, 2010 1:10 pm Post subject:

Computer, I have not written these types of programs but only supported them so you would probably need to verify what I am saying.