Re: De-duplicating (was Re: Finding duplicate rows in a table)

In article <tcoxCu2K2G.MH7_at_netcom.com> tcox_at_netcom.com "Thomas B. Cox" writes:

> >MarkB_at_aboy.demon.co.uk writes:
 

> >>The question is, are these rows truly identical throughout, or do they
> >>simply have identical primary keys?
>
> If he has "duplicate records" as he claims, then all values in all
> columns are the same for at least two rows.

Fair enough - I was trying to be exact in terms of defining the problem.
>
> >>If the former, it doesn't matter which you delete, therefore simply
> >>delete the one with the max(rowid) in a correlated query.
>
> Close, but not quite. Deleting using max(rowid) only takes out one
> row. What if you have three or four duplicates?

No - I don't think so. You can only have two in a pair of duplicates. If you have three rows the same, what you've got is a triplicate. The confusion arises because the word has taken on a different meaning (and in my opinion a misleading one, otherwise we wouldn't be discussing it) within RDBMS-speak than it has in the real English-speaking world.

Of course, within the context of this group, it was dumb of me to proceed without declaring my pedantry on such matters. My apologies.

Cheers,

M

-- 
-----------------------------------------------------------------------------
                                                    MarkB_at_aboy.demon.co.uk
Lyric Quiz of the Week: "You snatch a tune and you match a cigarette..."