Subject | RE: [IBO] Re: Cursor not updatable ??? |
---|---|
Author | Jason Wharton |
Post date | 2005-06-06T16:56:38Z |
I need a simple sample application demonstrating this problem to go any
farther.
Jason Wharton
www.ibobjects.com
farther.
Jason Wharton
www.ibobjects.com
> -----Original Message-----
> From: IBObjects@yahoogroups.com [mailto:IBObjects@yahoogroups.com]On
> Behalf Of Eyal
> Sent: Saturday, June 04, 2005 2:11 PM
> To: IBObjects@yahoogroups.com
> Subject: [IBO] Re: Cursor not updatable ???
>
>
> --- In IBObjects@yahoogroups.com, Helen Borrie <helebor@t...> wrote:
>
> > I suppose you do have RequestLive true?
>
> Yes.
>
>
> > If so, it seems as though your KeyLinks are not able to uniquely
> > identify an underlying row in the table you marked as KeyRelation.
>
> No. I've traced into IBO's code, and it generates a correct update
> statement, whether I have KeyLinksAutoDefine=TRUE or KeyLinks=table.id
> (the primary key column for the table).
>
> The problem is that IBO passes incorrect *value* for the search column
> - a value of zero, instead of the actual value of the id field.
>
>
> > Have you got an outer join there, with nulls falling into the
> > Keylinks?
>
> Outer join - yes, but KeyLinks point to a primary key column in the
> left (outer) table, which always has a non-null value.
>
> > Do you have enough elements in your KeyLinks to ensure uniqueness?
>
> Yes.
>
>
> > Show your SQL, KeyLinks and KeyRelation.
>
> SQL:
>
> SELECT acct.*, p.group_code
> FROM acct
> LEFT OUTER JOIN p_tree(1, 1, ''T'') p ON acct.id=p.obj_id
>
> KeyLinks = acct.id (the primary key for table acct).
> RequestLive = TRUE
> PreparedEdits = FALSE
>
> IBO correctly generates this update SQL:
>
> UPDATE acct SET name=? WHERE acct.id=?
>
> The problem is that the second value that IBO passes to the API call
> is always zero, instead of the actual id. I've verified it for a row
> with a known id (non-zero of course).
>
> Also, if I use my own EditSQL with exactly the same SQL (UPDATE acct
> SET name=:name WHERE id=:old_id) - it works just fine.
>
> So I guess the problem isn't the SQL, it's with the value that IBO
> passes to the API call when it auto-generates the update SQL.
>
> Regards,
>
> Eyal.