Question

This question is similar to what I need to do, but the voted answer only partially solves the problem. Consider this table:

ID   |  TEST  |  RESULT
001  |  AAA   |  +
001  |  BBB   |  +
002  |  AAA   |  +
003  |  BBB   |  +

I would like output like this:

ID   |  AAA |  BBB
001  |  +   |  +
002  |  +   |  
003  |      |  +

The solution I mentioned above provides the same data as its output, but does not “skip over” the empty columns, as for 003. Therefore, it is not possible to determine which test a + result refers to. I’m using PostgreSQL 8.4.9.

PS: Is a solution in pure SQL possible? Would this need to be coded in PL/pgSQL instead?

Answer 1

For a quick and dirty solution you can use this:

SELECT DISTINCT 
T.ID,
(SELECT A.RESULT FROM MyTable A WHERE A.TEST = 'AAA' AND A.ID = T.ID) AS AAA,
(SELECT B.RESULT FROM MyTable B WHERE B.TEST = 'BBB' AND B.ID = T.ID) AS BBB
FROM MyTable T;

If you need that more dynamic consider using crosstab function in the SELECT.