What the problem was is the locktype in the ado syntax was
adLockOptimistic and it was throwing the tablename.rowid into the sql
string. Why it didn't add the new column into the group by clause I'll
never know. So I switched the locktype to adLockReadOnly
and the extra rowid column was removed and the query stopped failing.
outerjoin wrote:
> Undergoing a conversion to Oracle on Linux ( redhat ) and am running
> into a baffling problem. During runtime an ADO query
>
> select col1,sum(col2) from table1 group by col1
>
> is generating a runtime error: invalid group by. I've found in SQL
> Monitor ( Quest freebie ) that the sql string it thinks is going to the
> server is
>
> select col1,sum(col2),col1.rowid from table1 group by col1
>
> Strangest thing is that the same code running against M$ works fine and
> has been
> running for 1+ years.
>
> We are having other issues but this is the problem that has been
> assigned to me.
>
> The application is running on server 2000 on Citrix 1.8 FWIW I'm
> experiencing the same
> problem running from the IDE on my laptop ( XP ), VB 6 SP 6.
>
> Thanks in advance for any help !
Received on Thu Jun 29 2006 - 12:07:01 CDT
