Create a query to compare values from the same table

Question

Suppose you have the below table, same ID's occur for same month as well as different month
ID Month Value
226220      201203     100
1660      201204     200
26739      201204     1010
7750     201205     31.1
I need a query to determine the below laid result
ID Month Prior_month_value Prior_Month Value
1234 201203 10 201201 100
3456 201206 56.1 201204 78
Please help
Edited by: Jaguar on Jul 10, 2012 3:00 AM

Answer 1

As mentioned, you can use the analytical function lag...
SQL> ed
Wrote file afiedt.buf
  1  with t as (select 226220 as id, 201203 as mnth, 100 val from dual union all
  2             select 1660, 201204, 200 from dual union all
  3             select 226220, 201204, 1010 from dual union all
  4             select 1660, 201206, 31.1 from dual union all
  5             select 7750, 201205, 60 from dual)
  6  --
  7  -- end of test data
  8  --
  9  select id
10        ,mnth
11        ,val
12        ,lag(mnth) over (partition by id order by mnth) as prev_month
13        ,lag(val) over (partition by id order by mnth) as prev_val
14  from t
15* order by id, mnth
SQL> /
        ID       MNTH        VAL PREV_MONTH   PREV_VAL
      1660     201204        200
      1660     201206       31.1     201204        200
      7750     201205         60
    226220     201203        100
    226220     201204       1010     201203        100Edit to Add...
and to filter out just the ones where there is a previous month...
SQL> ed
Wrote file afiedt.buf
  1  with t as (select 226220 as id, 201203 as mnth, 100 val from dual union all
  2             select 1660, 201204, 200 from dual union all
  3             select 226220, 201204, 1010 from dual union all
  4             select 1660, 201206, 31.1 from dual union all
  5             select 7750, 201205, 60 from dual)
  6  --
  7  -- end of test data
  8  --
  9  select id, mnth, val, prev_month, prev_val
10  from (
11        select id
12              ,mnth
13              ,val
14              ,lag(mnth) over (partition by id order by mnth) as prev_month
15              ,lag(val) over (partition by id order by mnth) as prev_val
16        from t
17       )
18  where prev_month is not null
19* order by id, mnth
SQL> /
        ID       MNTH        VAL PREV_MONTH   PREV_VAL
      1660     201206       31.1     201204        200
    226220     201204       1010     201203        100Edited by: BluShadow on 10-Jul-2012 11:32