Re: Records written not seen by called program

Once upon a time there was an issue with a program writing records to a file and another program opening the file to read wouldn't find the record.

IIRC, the solution was that a program that was only going to read the file should open it for full update. You had to include a dummy write to the file (in the old days it was a write statement conditioned by an indicator and its opposite (IE 01 N01)). This was to keep the compiler happy with a fully-functional file having to output to the file in question.

I thought that this error had been corrected long ago...

--Paul E Musselman
PaulMmn@xxxxxxxxxxxxxxxxxxxx

.


At 7:07 PM +0000 12/6/16, Roger Harman wrote:

I have a process that has failed a couple of times. PgmA writes a record and then calls PgmB with they key info. On rare occasions, the chain to the file from PgmB fails.

The file in question is not unique keyed and I'm wondering if the record is being buffered and not visible to the called program. I recall reading something about that a long time ago and that the fix is to have a unique key (LF?) over the physical file. Is that correct? Is FRCRATIO(1) essentially the same thing? It's not a high volume process. Maybe 12,000 records on a busy day.

Thoughts?

Thank you.
Roger Harman
COMMON Certified Application Developer - ILE RPG on IBM i on Power