On Oct 27, 2010, at 9:43 AM, Vick Khera wrote: > Any recommendations on how to fix? It is hard to trace the schema relations > since there are not FK's defined, so I'm not sure how to update the group id > references or if I need to. Also, is this table supposed to be really huge? > I have over 700k tuples in it!
In awesome form, I'll reply to myself ;-) Looks like there are two "groups" that are duplicate groupid+member id. It should be plenty easy to clean that up. What other table refers to the groupmembers.id field? I can just update those, delete the dupes, and have a nice unique index. Basically I did this: select groupid,memberid,count(*) from groupmembers group by groupid,memberid having count(*) > 1; It gave me this: groupid | memberid | count ---------+----------+------- 466380 | 134922 | 2 386385 | 134922 | 2
