Re: (un)grouping question

On Mon, 2008-01-21 at 12:36 -0500, Rhys Stewart wrote:

> uid|somevalue
> --------------------
> 1|11
> 2|44
> 3|31
> 4|44
> 5|71
> 6|33
> 7|33
> 8|44
> 9|14
>
> would like to remove the duplicate values in the column somevalue.
> doing this by just adding a random number is perfectly fine, however
> i want to retain at least one of the original values of somevalue. Any
> ideas how to do this in in a query?

There's certainly no need for a random number hack. Instead, use a query
like 'select distinct on (somevalue) * from mytable;' .
Although DISTINCT is standard, the DISTINCT ON (cols) variant is a
PostgreSQL-ism.

See
http://www.postgresql.org/docs/8.2/static/queries-select-lists.html#QUERIES-DISTINCT .

-Reece, your heterographic brother

--
Reece Hart, http://harts.net/reece/, GPG:0x25EC91A0