Thanks Joe that does the trick nicely.
Caroline

>Hi Caroline,
>
>How about something like this:
>
>SELECT a.iid, b.iid as iid__duplicate;
> FROM duplicates!addresses a, duplicates!addresses b;
> WHERE b.iid > a.iid;
> AND a.csurname == b.csurname;
> AND a.cforename == b.cforename;
> AND (a.caddress_1 == b.caddress_1;
> OR (a.cpostcode == b.cpostcode));
> ORDER BY a.csurname, a.cforename, a.iid
>
>This should result in:
>
>iid iid__duplicate
>
>1 2
>1 3
>2 3
>
>The key difference is using > vs. <>.
>
>When I use a similar approach, I assume that newer rows have a higher value for the UID, but I think it works regardless.
>
>HTH
>
>Blue Skies!
>
>Joe