Re: SQL question

If I read you right, you want the rows from Table_A based on the ID that do not exist in Table_B, and the Rows from Table_B based on the ID that do not exist in A?

I think this is what you want....
Try this:

• Records in A not in B SELECT ID FROM TABLE_A WHERE ID Not In (SELECT ID FROM TABLE_B) UNION ALL
• records in B not in A SELECT ID FROM TABLE_BWHERE ID Not In (SELECT ID FROM TABLE_A)

You may want to use DISTINCT clause in there too.

"Bastiaan Schaap" <bschaap_at_desyde.nl> wrote in message news:newscache$81z0zf$ypi$1_at_lnews.actcom.co.il...
> Hi all!
>
> I have a little query problem which I find very difficult to solve, I'm
 more
> of a plsql man ;-)
> Here it is: Say I have to tables A and B both with a column e.g. ID. Now I
> want to compare these tables. In order to do that I want to do a select
 that
> joins these two tables, and creates a result for every row in A and every
> row in B. Ergo: if A.ID exists and B.ID doesn't, I do want to see a row
 for
> it. Same is for B, if B.ID exists but A.ID doesn't I also want to see a
 row
> in the result set. The 'truth table' for this would like :
> A: B: row in result:
> 0 0 0 (does not occur, but I'll give it anyway..)
> 1 0 1
> 0 1 1
> 1 1 1
>
> Where a zero indicates that no row was found for the ID.
>
> Actually theoretically this works:
> SELECT A.ID as old_id, B.ID as new_id FROM A, B WHERE a.id(+) = b.id(+);
>
> However Oracle doesn't support this. Does anyone know a solution for this
> (little?) problem??
>
> Thanx,
>
> Bastiaan Schaap
> Oracle web development,
> Desyde BV - Baarn
> http://www.desyde.nl/
> tel. +31355411711
>
>
Received on Wed Aug 09 2000 - 00:00:00 CDT