Question

I have a table with an id and a name (an a bunch of other stuff not relevant for this query). Now I need an SQL statement that returns one row per distinct name and in that row I need the name and one id (can be any id).

The table is looking something like this:

id | name
---+-----
 1 | a2
 2 | a2
 3 | a4
 4 | a4
 5 | a2
 6 | a3

btw. using Postgres 8.4

Tried various combinations of grouping or joining with self. Is this even possible without creating extra tables?

Answer 1

Arbitrarily choosing to return the minimum id per name.

SELECT name, MIN(id)
    FROM YourTable
    GROUP BY name