Re:OVRDBF with INHWRT(*YES)

>Is there any gotchas that anyone is familiar with 
>concerning this keyword usage?

Let's say that you're trying to test a program without updating the data.
The program adds a billing record to a file if it doesn't already exist:

     C           KEY       CHAINBILLFILE             80 
     C   80                WRITEBILLFILE                

If you're doing INHWRT(*YES), none of the records will ever be written,
so if you have duplicate keys in your input data, this test will not 
catch them:

        INHWRT(*YES)    INHWRT(*NO)
KEY     CHAIN result            CHAIN result
1       Not found               Not found
2       Not found               Not found
1       Not found               Found: duplicate

As a result, when this runs in "test" mode, there are no duplicate keys 
found, but when this runs in "live" mode, there are.

Other places to watch for are where a program further down in the job
stream expects updated (or added!) records but doesn't get them.

The age-old rule still applies: know thy data flows!

Buck Calabro
Commsoft

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---