Question

I have a database formed like this:

 descr    | hours |   timer    
----------+-------+------------
 foo      |   1.3 | 2012-07-14
 foo      |   2.5 | 2012-07-15
 foo      |  2.35 | 2012-07-16
 bar      |     1 | 2012-08-16
 baz      |     1 | 2012-08-16
 buh      |     5 | 2012-08-17

This does what I expect:

abc=# select extract('month' from timer) as Month, sum(hours) from foo group by Month     order by 1;

 month | sum  
 ------+------
 7     | 6.15
 8     |    7

This doesn’t quite do what I expect:

abc=# select to_char(timer, 'Month'), sum(hours) from foo group by 1 order by 1 asc;

 to_char  | sum  
----------+------
August    |    7
July      | 6.15

The desired result is:

month     | sum
----------+-----
July      | 6.15
August    |    7

I understand why things are happening in #3, but how do I achieve the desired result, within PostgreSQL, in the most efficient manner possible?

Answer 1

You can use extract expression in ORDER BY section:

SELECT to_char(timer, 'Month'), SUM(hours)
FROM foo
GROUP BY to_char(timer, 'Month'), extract('Month' from timer)
ORDER BY extract('Month' from timer) ASC;

This query does desired result.