Readnext in Alternate index

cvishu · Active User Joined: 31 Jul 2007 Posts: 136 Location: india

Hi ,
I had a base cluster with key length as 20 , and two alteranate index for it.

I chnged the key length to 26 bytes in the base and adjusted the key position in the alternate index. i did all this through a Batch Delete-Define job .

But now when i do a REANEXT in the alternate index and check for End-Of-file , its just reading again and again indefinetly,without reaching EOF.
(The readnext used to work before the KEYS were changed)

Please advice if i should do any FCT entry change too ? or what else could be the issue with alternate index

PeterHolland · Posted: Wed Mar 17, 2010 3:10 pm

I think the Remote maximum key length (RMTKEYLEN)
in the FCT has to be changed too.

Robert Sample · Posted: Wed Mar 17, 2010 4:55 pm

You haven't provided enough information for us to help you. Are you doing a full key READNEXT or using a generic key? What release of CICS are you using -- FCT entries were replaced by RDO (resource definition online) many, many years ago? Are you using multiple regions (TOR and AOR for example) or is this being done in one region? Is the file assigned to an LSR pool? If so, what are the LSR pool characteristics?

And what does this mean?

cvishu · Active User Joined: 31 Jul 2007 Posts: 136 Location: india

Robert Sample · Posted: Wed Mar 17, 2010 6:18 pm

Use BBcode (the Code button) to post the EXACT code you are using to do the READNEXT -- from the EXEC CICS through the END-EXEC.

cvishu · Active User Joined: 31 Jul 2007 Posts: 136 Location: india

Robert Sample · Posted: Wed Mar 17, 2010 7:10 pm

cvishu · Active User Joined: 31 Jul 2007 Posts: 136 Location: india

Robert Sample · Posted: Wed Mar 17, 2010 8:06 pm

Many terms in IT have very precise definitions and using the wrong term in the wrong context (such as referring to a file having rows) at best indicates carelessness (or incompetence) on the person's part, and at worst indicates the person has absolutely no business being in IT in the first place. There have been quite a few posts on this forum where the person used the wrong term and cause a lot of misguided effort attempting to solve a problem that didn't exist while keeping responders from not focusing on the real issue. If someone isn't sure of the right term, and so indicates, that's a good way to learn along the way. But using the wrong term without knowing it implies much about the person ....

cvishu · Active User Joined: 31 Jul 2007 Posts: 136 Location: india

Robert ,

But my problem still persists. I have removed the Keylength from my READNEXT statement , still i dont see an y difference, the read is just going in loop without coming to EOF.

Robert Sample · Posted: Wed Mar 17, 2010 9:01 pm

Do you have a WHEN OTHER on your EVALUATE? There are a lot of possible responses on READNEXT other than NORMAL, NOTFND, and ENDFILE.

Earl Haigh · Active User Joined: 25 Jul 2006 Posts: 475

First:

run your transaction through an application debugger'
and capture the CICS reponse code you are receiving
from the Startbr and Readnext commands.

2nd:

If you can't resolve from that, Call your CICS technical
Support group and tell them your situation.

cvishu · Active User Joined: 31 Jul 2007 Posts: 136 Location: india

I did trace the module, and the response code for READNEXT is 0 .
And yes i do have a when other in my evaluate.

Robert Sample · Posted: Thu Mar 18, 2010 4:44 pm

First, I'm not quite sure why you would incur the overhead of the browse and READNEXT for a single record -- but you may have valid reasons to do so.

Second, at this point it appears to be an application issue. If you change the RIDFLD value in any way after the READNEXT, for example, CICS will reset the browse and return the first record found when it does the READNEXT -- so you could easily wind up in a loop.

To continue debugging this problem, I recommend you add a second record with the same alternate index key to the file and use CEDF to trace through the code to see what happens.