Re: SQL Cursors and *ENDMOD

[CRPence <CRPbottle@xxxxxxxxx>]

Alan Campin wrote:
> I am having problems with clears being able to clear members and
> I have traced the problem back to the SQL statements in a service
>  program.

  If the cursor is supposed to be closed but the lock & open remain 
on the file.mbr, then it may be pseudo-closed cursor.  If so, then 
if the request to WRKJOB OPTION(*RUNA) /* or is it *DFNA? */ shows a 
/small/ Default Wait value, then try CLRPFM after CHGJOB DFTWAIT(30) 
to give time for the holding job(s) to effect their close activity.

> It appears that I was not closing an SQL cursor after I was done
> with it.

  Even a CLOSE will not ensure avoiding a pseudo-close by the SQL 
database feature.  The SQL might choose to pseudo-close instead of 
full-close, with the expectation the same cursor is simply going to 
be opened again; e.g. as when open\close pair is in a coded loop.

> <<SNIP *ENDMOD stuff>>
>
> Now the fun part is that when I run it in test I don't see the
> cursor open after it returns.

  The open/close cycle must be repeated for the statement to have 
the pseudo-close feature activate.

Regards, Chuck