Query problem with multiple date fields.

Question

I need to select the greatest date from six different fields in each record and identify from which column was selected.
date01 date02 date03 date04 date05 date06
010190 010107 010190 010190 010190 010190
010190 010190 010105 010190 010190 010190
010103 010190 010190 010190 010190 010190
1)if greatest is in date01 then S = 1
2)if greatest is in date02 then S = 2
3)if greatest is in date03 then S = 3
4)if greatest is in date04 then S = 4
5)if greatest is in date05 then S = 5
6)if greatest is in date06 then S = 5
Do you have any idea on how to acomplish this?
Thanks

Answer 1

Another way (won't work for null dates):
with
the_data as
(select to_date('01011990','DDMMYYYY') col1,to_date('01012007','DDMMYYYY') col2,to_date('01011990','DDMMYYYY') col3,
        to_date('01011990','DDMMYYYY') col4,to_date('01011990','DDMMYYYY') col5,to_date('01011990','DDMMYYYY') col6
    from dual union all
select to_date('01011990','DDMMYYYY'),to_date('01011990','DDMMYYYY'),to_date('01012005','DDMMYYYY'),
        to_date('01011990','DDMMYYYY'),to_date('01011990','DDMMYYYY'),to_date('01012005','DDMMYYYY')
    from dual union all
select to_date('01012003','DDMMYYYY'),to_date('01011990','DDMMYYYY'),to_date('01012003','DDMMYYYY'),
        to_date('01011990','DDMMYYYY'),to_date('01012003','DDMMYYYY'),to_date('01011990','DDMMYYYY')
    from dual
the_setup as
(select rownum the_row,col1,col2,col3,col4,col5,col6,greatest(col1,col2,col3,col4,col5,col6) the_greatest,
        to_char(col1,'DDMMYYYY') || to_char(col2,'DDMMYYYY') || to_char(col3,'DDMMYYYY') ||
        to_char(col4,'DDMMYYYY') || to_char(col5,'DDMMYYYY') || to_char(col6,'DDMMYYYY') dates
   from the_data
the_columns as
(select the_row,
        max(decode(position,1,'1,')) || max(decode(position,2,'2,')) || max(decode(position,3,'3,')) ||
        max(decode(position,4,'4,')) || max(decode(position,5,'5,')) || max(decode(position,6,'6,')) the_cols
   from (select distinct the_row,level position
           from the_setup
          where substr(dates,8 * (level - 1) + 1,8) = the_greatest
         connect by level <= 6
group by the_row
select col1,col2,col3,col4,col5,col6,the_greatest,rtrim(the_cols,',') the_columns
  from the_setup s,the_columns c
where s.the_row = c.the_row
order by s.the_row
==========================================================================================================
COL1       | COL2       | COL3       | COL4       | COL5       | COL6       | THE_GREATEST | THE_COLUMNS
==========================================================================================================
01.01.1990 | 01.01.2007 | 01.01.1990 | 01.01.1990 | 01.01.1990 | 01.01.1990 | 01.01.2007   | 2
01.01.1990 | 01.01.1990 | 01.01.2005 | 01.01.1990 | 01.01.1990 | 01.01.2005 | 01.01.2005   | 3,6
01.01.2003 | 01.01.1990 | 01.01.2003 | 01.01.1990 | 01.01.2003 | 01.01.1990 | 01.01.2003   | 1,3,5Regards
Etbin