>Hello,
>
>I have a table that keeps information in a kind of sub-records of other records in the same table.
>Each record has an IDKey and a RelationKey. In the RelationKey the number of the parent record (in the same table!) is stored, if the RelationKey = 0 the record indicates a parent node without having other parents above. (I hope I have successfully explained this!)
>
>So the data could look like:
>

>IDKey   RelationKey   Success
>   1             0     Y
>   2             1     N
>   3             1     Y
>   4             2     y
>   5             4     Y
>

>Now I need to select all records that have the field "Success" = "Y", but ONLY those where also ALL parents have Success = "Y".
>
>I would like to make an inner join or so, but the problem is that I don't know how many subrecords will be there. In this case, record 5 should not be selected, although record 4 (the parent of 5) is set to "Y", because record 2 (the parent of 4) is set to "N".
>How can I solve this problem?

I do not think there is a solution to this because you can have an unlimited recursion up the parent tree. If you changed the success to Y for IDKey 2 and N for IDKey 1, you still would not select record 5. If you were only going back 1 level [the parent and no grandparents], this would be fairly easy.