SEARCH ALL - WHEN statement ignored

jtwohig · New User Joined: 24 Jul 2006 Posts: 18

Hi,

I have a table I'm doing a SEARCH ALL on and can't figure out why
the "WHEN" statement is apparently being ignored. Table is below.
I've put displays in prior to the WHEN statement and both the fields
in the WHEN stmt are populated and are a match. I put a display stmt
after the WHEN but it (display) doesn't show up. Been looking at this for
a while, I'm stumped.

Thanks,
J.

dick scherrer · Posted: Tue Mar 13, 2007 2:22 am

Hello,

How have you verified that the table is loaded and is in the proper sequence?

Before working on your "SEARCH ALL", you might want to display the contents of the table from TABLE1-PERIODS(1) through whatever is the max for this run. What is in WT-TABLE-CNT?

Also, when you post your code, it is helpful is you post using the "Code" tag at the top of the replay panel.

jtwohig · New User Joined: 24 Jul 2006 Posts: 18

Thanks Dick !
I had put displays of the table field just prior to the WHEN statement and it showed it was populated. I didn't do a display, per se, of all of the table,I can try that. WT-TABLE-CNT holds the total recs loaded to the table. Will use "code" tag in future.
Thanks,
J.

William Thompson · Posted: Tue Mar 13, 2007 2:38 am

One key out of sequence will definitely mess things up, it is obvious that the search is "at end"ing prior to finding a match. Confirm the OCO counter and the sequence of the keys. also confirm the key and the search argument sizes are the same.

dick scherrer · Posted: Tue Mar 13, 2007 3:46 am

Hi J,

You're welcome

Just for testing, you might use SEARCH instead of SEARCH ALL. You might also set the index to 1 and go thru the table to verify if "this" TABLE1-PROV-TYPE is greater than the "previous" value. One out of sequence entry may cause the search to fail.

This may just be picky, but if the array can occur 500 times, i'd make the WT-TABLE-CNT (3) COMP-3 VALUE 0.

When i asked about WT-TABLE-CNT, i was looking for the current value. How many entries are there for this test?

You could look at the code that loads the table. It should "sequence check" as it loads the array.

Good luck - Waiting to hear from you

jtwohig · New User Joined: 24 Jul 2006 Posts: 18

Thanks William and Dick,

Let me check out some of the things you suggest.

Dick, you mentioned checking the code that loads the table. I'm not sure I know what you mean by a "sequence check" ?

Thanks,
J.

dick scherrer · Posted: Tue Mar 13, 2007 8:26 am

Hi J,

By that i mean to have some code that makes sure that each entry moved into the array has a "key" that is higher than the preceeding entry. For example, if there is a "GG" prov-type to be moved into the array, the previous entry must be "GF" or lower. Every entry to the table must be greater than the one before.

If you post the code that moves values to the array, it will be helpful.

jtwohig · New User Joined: 24 Jul 2006 Posts: 18

Thanks Dick. I just now discovered my problem and I'm afraid it was a stupid one I can't believe I missed. In the table's "DEPENDING ON" phrase, I have "DEPENDING ON WT-TABLE-CNT". Well, in cloning code from another program I loaded the table adding to a field "WA-TABLE-CNT", so, naturally, WT-TABLE-CNT was always zero. Just shoot me !
Thanks again for your help.
J.

William Thompson · Posted: Tue Mar 13, 2007 5:34 pm

jtwohig · New User Joined: 24 Jul 2006 Posts: 18

Thanks again William. Unfortunately, I'm sure this won't be the last time I'm chasing my tail...

dick scherrer · Posted: Tue Mar 13, 2007 7:57 pm

Hi J,

I hate when i do that. . . .

Glad you "caught" the tail. You got caught in one of my least favorite situations - the compiler believed what it saw - would have been better if the field-name did not exist (if only there was a "do what i meant" compile parameter). But think of all the "fun" you had

Good luck and be in touch.