SQL query to correct the corrupted data

Question

Hi All,
I got strucked up with a query correcting the corrupted records.
CREATE TABLE EX1
EMPID INTEGER,
DW_EFF_DT DATE,
DW_EXPR_DT DATE
INSERT INTO EX1 VALUES(1,'04-MAR-1998','13-MAR-1999');
INSERT INTO EX1 VALUES(1,'14-MAR-1999','02-MAY-2000');
INSERT INTO EX1 VALUES(1,'03-MAY-2000','01-MAY-2013');
INSERT INTO EX1 VALUES(1,'02-MAY-2013','31-DEC-9999');
I have empid with other attributes,we are maintaing history.There is some data which is corrupted and we need to correct it.
DW_EFF_DT which is less than '01-FEB-2005' should be made by default '31-JAN-2005'
Again the data should be corrected and my output should be like the below
EMPNO     DW_EFF_DT     DW_EXPR_DT
1     2005-01-27     2005-01-28
1     2005-01-29     2005-01-30
1     2005-01-31     2013-05-01
1     2013-05-02     9999-12-31
I used the lead and lag function,but it is applied sequentially.
How I can get the dates subtratcing by 1 for each and every alternative row.
I tried the below query and able to achive half of it.
SELECT A.* ,COALESCE(LEAD(NEW_DW_EFF,1) OVER(ORDER BY NEW_DW_EFF) -1,TO_DATE('31-DEC-9999','DD-MON-YYYY')) AS NEW_DW_EXPR_DT
FROM
SELECT ID,
DW_EFF_DT,
DW_EXPR_DT,
CASE WHEN DW_EFF_DT<TO_DATE('01-FEB-2005','DD-MON-YYYY') THEN TO_DATE('31-JAN-2005','DD-MON-YYYY') ELSE DW_EFF_DT END AS NEW_DW_EFF
FROM EX1
)A
ID     DW_EFF_DT     DW_EXPR_DT     NEW_DW_EFF     NEW_DW_EXPR_DT
1     04-MAR-98     13-MAR-99      31-JAN-05      30-JAN-05
1     14-MAR-99     02-MAY-00      31-JAN-05      30-JAN-05
1     03-MAY-00      01-MAY-13      31-JAN-05      01-MAY-13
1     02-MAY-13      31-DEC-99      02-MAY-13      31-DEC-99
Please help me in this regard.i am still trying it.
Thanks in advance,
KVB

Answer 1

KVB wrote:
EMPNO     DW_EFF_DT     DW_EXPR_DT
1     2005-01-27     2005-01-28
1     2005-01-29     2005-01-30
1     2005-01-31     2013-05-01
1     2013-05-02     9999-12-31
Actually the last record is the active record.
Th 3rd record effectiuve date is adjusted because the dateis less than 2005-02-01. So that we adjusted to 2005-01-31,
The 2nd record expiry date will be adjusted to 3rd record effective date-1,
Now the 2nd record effective will be adjusted to 2nd record expiry date-1 and so on..until the first record.
Actually I could not explain much better in business point of view like why they do this.Technically we need to achive this through SQL.
CheersOkay, that makes more sense, thanks :)
Probably not the prettiest ... but should work (you may need to tweak it a bit)
ME_TUBBZ? select
  2        empid
  3     ,  to_date('2005-02-01','yyyy-mm-dd')  - ( 2* rn) + 1 as new_eff_date
  4     ,  case
  5           when rn = 1
  6           then
  7              dw_expr_dt
  8           else
  9              to_date('2005-02-01','yyyy-mm-dd') - (2 * (rn - 1))
10        end
11              as new_exp_date
12     ,  dw_eff_dt
13     ,  dw_expr_dt
14  from
15  (
16     select
17           empid
18        ,  dw_eff_dt
19        ,  dw_expr_dt
20        ,  row_number() over (partition by empid order by dw_eff_dt desc)             as rn
21     from ex1
22     where dw_eff_dt < to_date('2005-02-01','yyyy-mm-dd')
23  )
24  order by dw_eff_dt  asc
25  /
             EMPID NEW_EFF_DATE         NEW_EXP_DATE         DW_EFF_DT            DW_EXPR_DT
                 1 27-JAN-2005 00 00:00 28-JAN-2005 00 00:00 04-MAR-1998 00 00:00 13-MAR-1999 00 00:00
                 1 29-JAN-2005 00 00:00 30-JAN-2005 00 00:00 14-MAR-1999 00 00:00 02-MAY-2000 00 00:00
                 1 31-JAN-2005 00 00:00 01-MAY-2013 00 00:00 03-MAY-2000 00 00:00 01-MAY-2013 00 00:00
3 rows selected.
Elapsed: 00:00:00.01
ME_TUBBZ? Cheers,