2008/10/7 Yi Zhao <yi.zhao@alibaba-inc.com>:
> I have a table contains some duplicate records, and this table create
> without oids, for example:
> id | temp_id
> ----+---------
> 10 | 1
> 10 | 1
> 10 | 1
> 20 | 4
> 20 | 4
> 30 | 5
> 30 | 5
> I want get the duplicated records removed and only one is reserved, so
> the results is:
> 10 1
> 20 4
> 30 5
>
> I know create a temp table will resolve this problem, but I don't want
> this way:)
>
> can someone tell me a simple methold?
>
> any help is appreciated,
>
> thanks,
>
>
I would not say this is easier....
1. alter table t add key serial;
2. delete from table where key not in (select max(key) from table
group on id,temp_id);
The truth is this is not any less work then using a temporary table
(whole table still needs rewriting). Which method you select really
depends on why these duplicate records exist in the first place.
Regards
Peter