SV: [V4RB] insert into select quirk

[niklas wormann]

>
>
> on 3/3/04 12:36 AM, niklas wormann at nwoe at privat.utfors.se wrote:
>
> > Hi!
> > 1.10 version
> >
> > I am using the famous workaround of method fields for compound indexes.
> > created a method field like concat(nodeid,'#',caseid) and set
> field flag to
> > unique
> >
> > when using the "insert into select" syntax i get no info about one of
> > my records breaking the laws of my compound index.
> > error is 0, errstring is ok and recordcount is as expected.
> > BUT  no records are inserted.
> >
> > would be nice to get some kind of message about this.....
>
> Agree Niklas,
>
> Probably no error because of this is batch mode...
>
> So you do INSERT INTO table that have BaseObject method?
> Of FROM this table?
>
> If you try write INTO BaseObject method..then
> I think Valentina should ignore your values, and insert into other real
> fields.
>

i am inserting into table that has baseobject method.
because of poor design by me of method field
concat(nodeid,caseid)-- gives duplicate values for e.g nodeid 11, caseid 1
and nodeid1, caseid 11.


so when i try to insert,i am not trying to insert inte the method (!), into
caseid and nodeid,
i dont get any errors, although the insert fails completely. no records are
inserted.

everything seems to work out ok from err and errstring, but it isn?t.

niklas