There have been no end of problems with 7.6.2 xref segments. In my case these are DKU's.

I have just checked and found out that IBI have fundamentally changed the structure of the parsed master file table, SYSCOLUM, in this release playing around with the SEGMENT table.

As Erik von Daniken would say: A co-incidence? I sink not!

Here are the problems

1. DKUs using ALIAS es no longer work
More Explanation Here

2. DKUs show the CRSEGNAME and not the SEGNAME in Report Assistant

3. In addition programs where there are DKU segments on the MASTER that crash under 7.6.2 work once you remove the DKUs .

I think I have identified the root cause of this problem. The SYSCOLUM parsed version of MASTER files has been modified fundamentally in rel 7.6.2.
Now, instead of a TABLE FIELD ACCEPT structure, a fourth segment has been added so the structure is now TABLE SEGMENT FIELD ACCEPT. The problem is that DKU segments are not properly being mapped to the structure - the CRSEGNAME is being used instead of the SEGNAME. This violates the premise that SEGNAME is UNIQUE in the structure and thus many problems emanate from this.

This reproduces the problem

-* File dkuproblem.fex
-GOTO STEP1
-STEP1
-* Get the data from CAR file to build XREF structure.
APP HOLD BASEAPP
TABLE FILE CAR
SUM COUNTRY 
BY CAR
ON TABLE HOLD AS CARXREF FORMAT XFOCUS INDEX CAR 
END
-RUN
TABLE FILE CAR
PRINT
     BODYTYPE
BY CAR
BY MODEL
ON TABLE HOLD AS CARHOST FORMAT XFOCUS
END
-EXIT
-STEP2
-* Now edit CARHOST.MAS and include the cross reference at the end of the MASTER  
-* SEGMENT=CAR_EXTEND,SEGTYPE=DKU,CRFILE=CARXREF,CRSEGNAME=SEG01,CRKEY=CAR,$
-EXIT
-STEP3
-* check it all works 
TABLE FILE CARHOST
PRINT *
END
-EXIT
-STEP4
-* now look at the SYSCOLUM that is created at SEGMENT level  
-* you will see that the SEGNAME is SEG01 whereas the LOCSEGNAME has the correct segment name of SEG01 AND CAR_EXTEND
TABLE FILE SYSCOLUM
SUM SEGNAME LOCSEGNAME 
BY SEGNO 
IF TBNAME EQ 'carhost'
END
-EXIT
-STEP5
-* now use Report Assistant against this MASTER - the incorrect SEGNAME is showing.

I have raised this Case number 32402551, but please can anyone let me know if this a known problem or if they have similar problems because its getting to where I'm tearing my hair out.

This message has been edited. Last edited by: hammo1j,

After reading your chronicles I'm surprised you still have hair - patience, yes, hair, no.

Francis.

After playing around trying to replicate the problem with CAR file (could not get really simple example) I would assume that the problem is a bit more obscure than I first thought.

I have managed to replicate the case but the example is relatively complex. It seems to be the problem arises when you attempt to use an alternate view (TABLE FILE master.fieldname) against a MASTER that contains DKU segments.

The case is 2849.01/32422508

Just so everyone knows. There is problem with DKU segments, alternate file views and WHERE tests. However, that can be worked around with SET WHOPT = OFF, until I get the fix from the programmers. . However, the later problem in this case so far we have not been able to reproduce. (alternate file view giving index not found) As soon as we can do so, I'll post up my findings here, so that everyone can see. Working on it, just as he says fairly obscure, DKU's , alternate file views, redefined fields on the DKU segment, etc.
I'll let you guys know. Noreen

WEFOCUS - All releases starting at 4.3.6
MF FOCUS -- All releases starting at 3.8.68
OS/Platfor _ Predominately Z/OS, but all others as swell
Expected Output Format -- *