sql - Postgresql SELECT DISTINCT ON -

in query below i'm trying select data need distinct on columns book.title , orderdate. have tried using distinct on multiple results post below. sorry if question bit trivial. i'm in process of learning sql i'm bit of noob when comes obvious have missed.

how change query results formatted so:

   month   |                    title                    | quantity | total_value  -----------+---------------------------------------------+----------+-------------  february  | internet , world wide web: how program |       15 |      899.70  march     | c how program                            |        4 |      183.92  march     | core servlets , javaserver pages          |       13 |      856.70  march     | internet , world wide web: how program |       21 |     1071.58 

the query constructed:

select distinct on (orderdate, book.title) book.title, to_char(orderdate, 'month') "order date",  orderline.quantity "order quantity", (sum(quantity*unitsellingprice)) "total value" book inner join publisher on book.publisherid=publisher.publisherid inner join orderline on book.bookid=orderline.bookid inner join shoporder on orderline.shoporderid=shoporder.shoporderid publisher.name='prentice hall' group book.title, orderdate, orderline.quantity order orderdate asc; 

the results get:

                    title                    | order date | order quantity | total value  ---------------------------------------------+------------+----------------+-------------  internet , world wide web: how program | february   |             10 |      299.90  internet , world wide web: how program | february   |              5 |      149.95  c how program                            | march      |              3 |       68.97  core servlets , javaserver pages          | march      |             10 |      329.50  internet , world wide web: how program | march      |             20 |      519.80  c how program                            | march      |              1 |       22.99  core servlets , javaserver pages          | march      |              3 |       98.85  internet , world wide web: how program | march      |              1 |       15.99 

thanks help!

i think you're over-complicating things here: if want total quantity each title , month, want group by title , month, , sum quantity. don't want group by quantity itself, because mean new row each distinct quantity.

that's easy in standard sql (distinct on postgres extension, useful in few cases instead of group by, unnecessary here):

select     book.title,     to_char(orderdate, 'month') "order date",      sum(orderline.quantity) "order quantity",      sum(quantity*unitsellingprice) "total value" [...] publisher.name='prentice hall' group     book.title,     to_char(orderdate, 'month') order to_char(orderdate, 'month') asc; 

the rule of thumb in select clause - , order by clause, considered "at same level" - should either:

• listed in group by clause because want new row every time varies
• or wrapped in aggregate function (here sum()) because want "roll up" according rule.