Tried that, it's zero.  Here is the code fragment;

  01 WS-ARG               PIC X(02)  VALUE "@".


     %DBFIND(BR-DB#,DS-STATE-M#,MODE21#,BR-STATUS-AREA#
             ,DI-STATE#,WS-ARG#).
      IF NOT BR-CALL-OK
         %DBEXPLAIN(2#,BR-STATUS-AREA#)
         STOP RUN.
      %DBGET(BR-DB#,DS-STATE-M#,MODE5,BR-STATUS-AREA#,STATE-M#
            ,WS-ARG#).
      PERFORM UNTIL NOT BR-CALL-OK
         DISPLAY STATE-M
         %DBGET(BR-DB#,DS-STATE-M#,MODE5,BR-STATUS-AREA#,STATE-M#
               ,WS-ARG#)
      END-PERFORM.

There are 51 entries in the set, and the program will display 51 blank
lines.

shawn





John Zoltak <[log in to unmask]> on 09/03/98 07:51:57 AM

To:   'Shawn Gordon' <[log in to unmask]>, 'HP3000-L'
      <[log in to unmask]>
cc:    (bcc: Shawn Gordon/IS/FHM/FHS)
Subject:  RE: [HP3000-L] More on image b-tree problem




Shawn,

Check the transfer count of the DBGET mode5, in the status array, word
2, and see if any data was actually transferred and maybe it was placed
somewhere other than you expected.

John Zoltak
North American Mfg Co

> -----Original Message-----
> From:   Shawn Gordon [SMTP:[log in to unmask]]
> Sent:   Thursday, September 03, 1998 10:33 AM
> To:     [log in to unmask]
> Subject:     [HP3000-L] More on image b-tree problem
>
> I ran this through a debugger and what I see is that my DBFIND mode 1
> on
> the master returns the correct chain count.  Then each DBGET mode 5 on
> the
> master is successfull, but the buffer that is returned is empty.  It
> will
> happily read all the records in the super chain until it hits end of
> chain,
> but there is never any data in the buffer.  I am using an @ for the
> item
> list, and an @ for the argument to the DBFIND.  Everything looks fine,
> but
> the DBGET never has any data in it.
>
> I have tried this with both modes 1 and 21 on the DBFIND, just like
> QUERY.
>
> Is this a bug in the b-trees or am I doing something really dumb?
>
> thanks,
>
> shawn