My suggestion is to use the SQL trace monitor to carefully examine exactly
what is being sent to the server. You may find some clues there.

Jason

> -----Original Message-----
> From: IBObjects@yahoogroups.com [mailto:IBObjects@yahoogroups.com]On
> Behalf Of Chuck Belanger
> Sent: Tuesday, November 27, 2007 10:29 AM
> To: IBObjects@yahoogroups.com
> Subject: Re: [IBO] Update and IB_DSQL what am I doing wrong?
>
>
> Thanks, Alan:
>
> Like I said maybe I'm missing something very basic, so I
> don't mind at all.
>
> What I mean is that I'm updating, i.e. changing a single
> small int field
> on a single record with a specific PK. Then I do that again,
> through a
> user initiated event--a popup menu click-- to another single record,
> with a completely different PK and the first record loses it changes,
> but the second (assuming I don't do a 3rd record update) remains. My
> question is how do I keep that first record changed, then apply (i.e.
> reissue an update for another record) the second change, and so on?
>
> The various usual methods are not working. What I finally found that
> kind of works (although still not acceptable) is to have the
> user select
> all items (tree nodes) first then apply a looped batch update for all
> items. The first time of applying this batch, it "holds", all
> successive
> batch updates seem to miss a single item, and in a batch
> update it looks
> like its consistently the LAST item updated which is missed, which is
> opposite to doing single updates, serially, in which the LAST
> item remains.
>
> What I find scary about this is how many other places in my
> code is this
> happening, too, and I just never looked close enough at the
> data to see it?
>
> Finally, it doesn't matter what method I use to "Update", the
> result is
> the same. (e.g. using a SELECT, EDIT, POST sequence).
>
> Thanks, I appreciate any help.
>
> Chuck
>
> Alan McDonald wrote:
> > Forgive me for butting in Chuck.
> > You say "only the last update actually takes"
> > I have to query the use of language here....
> > Obviously an update statement overwrites previous
> updates.... So the last
> > update is always the only one to remain "visible"
> > Can you comment on this observation please?
> > Alan