sql - Sort result when using ROW_NUMBER() -

this follow-up this question asked yesterday.
did not know if should expand original question, decided start new 1 (bear me...)

my select @ moment:

select * (     select          a.guid_column1, b.guidid_column1, c.date_column1, d.guid_column1         row_number() on (partition a.guid_column1, b.guid_column1 order c.date_column1 desc) rn      ...     -- joins , stuff here     .... ) t t.rn = 1 

i (expected) result this:

a.guid_column1   b.guid_column1    c.date_column1    d.guid_column1 ------------------------------------------------------------------- a1               b1                07/08/2013       someuniqueid a2               b2                05/06/2012       someuniqueid 

the tricky part sort result dateadd, order dateadd(month, a.float_column, c.date_column1) asc
a.float_column of course not same (and in fact entered user later on).

is there way accomplish in sql (i'm using sql server 2005)

below query calculate new date in select-clause.

if a.float_column null nothing added original date.

select dateadd(month, isnull(t.float_column,0), t.date_column1) newdate     , * (     select          a.guid_column1, b.guidid_column1, c.date_column1, d.guid_column1, a.float_column         row_number() on (partition a.guid_column1, b.guid_column1 order c.date_column1 desc) rn      ...     -- joins , stuff here     .... ) t t.rn = 1  order newdate asc