RE: SQL vs V5R1 and service woes

Thank you for the many suggestions, and to the individual who sent me on to
that one expert ;-)

So far I've done the following:
1)  Rewritten the entire program into free format SQLRPGLE.
2)  Tried the following SQL statements:
     C*EXEC SQL
     C* insert into plline2
     C*   (select * from vdline2
     C*    where (vdlpo concat h2item)
     C*          in (select (pdlpo concat plitem) from plline))
     C*END-EXEC

     C*EXEC SQL
     C* insert into plline2
     C*  (select * from vdline2 a
     C*    where exists (select * from plline b
     C*          where a.vdlpo = b.pdlpo and a.h2item = b.plitem))
     C*END-EXEC

     C/EXEC SQL
     C+ insert into plline2
     C+   (select * from vdline2
     C+    where (vdlpo concat h2item)
     C+    in (SELECT  a.vdlpo  concat  a.h2item
     C+    FROM  vdline2  a  INNER JOIN  plline  b
     C+    ON  a.vdlpo  concat  a.h2item  =  b.pdlpo  concat  b.plitem))
     C/END-EXEC

And still, the lines will not work in imbedded SQL.  If I run this program
in debug, break right after the failing statement, and run the SQL
interactively, it will work.  Now I am reaching for some real straws:
a)  The SQL before it adds the rows to the PLLINE file.  Does this need
some time to simmer?  Is this why the interactive works?
b)  Could it be some activation group issue affecting the difference
between imbedded and interactive?
c)  The devils in database trying to get back at those RPG people for
advancing the compiler faster than they can keep up with a simple, lousy,
precompiler?


And still, I can restore this to our V4R5 machine and it will work
flawlessly.

Rob Berendt

==================
A smart person learns from their mistakes,
but a wise person learns from OTHER peoples mistakes.