Question

I have table as

P_Id | userid | year | month | day
-----+--------+------+-------+------
 3   |      3 | 2011 |     2 |    2
 5   |      1 | 2011 |     2 |    3
16   |      8 | 2011 |     3 |    4
 5   |      3 | 2011 |     4 |    4
17   |     1  | 2011 |     4 |    6
 8   |      4 | 2011 |     7 |    7
 9   |      3 | 2011 |     8 |    8
10   |      8 | 2011 |     9 |    9

I want to select distinct column i.e userid but also the respective value of year month and year which were encountered first.

For given above table following should be output

P_Id | userid | year | month | day
-----+--------+------+-------+------
   3 |      3 | 2011 |     2 |    2
   5 |      1 | 2011 |     2 |    3
  16 |      8 | 2011 |     3 |    4
   8 |      4 | 2011 |     7 |    7

or
If i am ordering the table by year,month and day
userid which is encountered first must only be selected and rest must be not be selected

Answer 1

Put year, month and day to native date column and do this:

select p_id, userid, min(the_date) from table group by p_id, userid

It will provide the fastest result.

If you cant modify your table and should use year+month+day then you can convert this values to date and still use min function.