Re: Object Allocation Problem - or Something Else?

[Charles Wilt <charles.wilt@xxxxxxxxx>]

Ah...no.

COMMIT(*NONE) just means you're defaulting to not using commitment
control for any of the statements in the source used by the RUNSQLSTM.
 If you were doing INSERTs or UPDATEs, you could still explicity use
commitment control for a particular statement by adding the WITH CHG
clause.

However, on the i, commitment control is only use on SQL DML (data
manipulation lanaguage, ie. SELECT, UPDATE, DELETE, INSERT) not on SQL
DDL.

Even if COMMIT(*NONE) affected DDL (as it does on some platforms), it
would just affect the creation of the table, not the functionality of
the table.

The only thing that effects rather or not you can use commitment
control when doing I/O on a table/physical file is the fact that the
table must be journaled in order to use commitment control when doing
I/O to it.

HTH,
Charles

On Tue, Feb 21, 2012 at 2:44 PM, Jerry C. Adams <midrange@xxxxxxxx> wrote:
> My program, which processes the file, is simply RPG IV; no SQL in it.  I
> meant that, when I created the DTOORDERS table, I used DDL and ran RUNSQLSTM
> with COMMIT(*NONE).
>
> Commitment control is, admittedly, one of the things I have *never* used so
> I may be confusing both of us.  Anyway, I thought that, if the table was
> created with COMMIT(other than *None), that the program had to use COMMIT to
> actually add/update/delete the records/transactions.
>
> And, if the table was created with COMMIT(*None) that you couldn't use
> COMMIT in the program (or that it wouldn't have any effect, at least).