Read multiple rows to do aggregation

Question

Hi All
I need some help here:
1. I have to read multiple rows to check where time stamp is same for example rownum (1-3) : *'2013/03/09 10:54:09 PM'*
2. And status is unique for example *(H,L)*
3. And if TIME_STAMP is between TIME_A and TIME_B then flag it as Y else flag it as N.
and if the above condition is not true then
1. Flag the row with ' Y ' where time_stamp is same i.e. : '2013/03/09 10:54:09 PM'
2. And the status is *( ' O ')*
3. the other rows where time_stamp is same, and unique_id is same, and STATUS is either or both ( ' H ' , ' L ' ) flag it as *' N '*
table structure:
CREATE TABLE T2
  TIME_STAMP  DATE,
  STATUS      VARCHAR2(1 BYTE),
  UNIQUE_ID   NUMBER,
  TIME_A      DATE,
  TIME_B      DATE
Insert statement;
SET DEFINE OFF;
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 22:54:09', 'MM/DD/YYYY HH24:MI:SS'), 'H', 6587797, TO_DATE('03/09/2013 23:00:22', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 23:03:00', 'MM/DD/YYYY HH24:MI:SS'));
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 22:54:09', 'MM/DD/YYYY HH24:MI:SS'), 'L', 6587797, TO_DATE('03/09/2013 22:48:35', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 23:00:22', 'MM/DD/YYYY HH24:MI:SS'));
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 22:54:09', 'MM/DD/YYYY HH24:MI:SS'), 'O', 6587797, TO_DATE('03/09/2013 23:03:00', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 23:05:54', 'MM/DD/YYYY HH24:MI:SS'));
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 22:54:57', 'MM/DD/YYYY HH24:MI:SS'), 'H', 6587797, TO_DATE('03/09/2013 23:00:22', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 23:03:00', 'MM/DD/YYYY HH24:MI:SS'));
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 22:54:57', 'MM/DD/YYYY HH24:MI:SS'), 'L', 6587797, TO_DATE('03/09/2013 22:48:35', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 23:00:22', 'MM/DD/YYYY HH24:MI:SS'));
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 22:54:57', 'MM/DD/YYYY HH24:MI:SS'), 'O', 6587797, TO_DATE('03/09/2013 23:03:00', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 23:05:54', 'MM/DD/YYYY HH24:MI:SS'));
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 18:26:55', 'MM/DD/YYYY HH24:MI:SS'), 'L', 6583483, TO_DATE('03/09/2013 18:15:51', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 18:26:28', 'MM/DD/YYYY HH24:MI:SS'));
Insert into T2
   (TIME_STAMP, STATUS, UNIQUE_ID, TIME_A, TIME_B)
Values
   (TO_DATE('03/09/2013 18:26:55', 'MM/DD/YYYY HH24:MI:SS'), 'O', 6583483, TO_DATE('03/09/2013 18:27:01', 'MM/DD/YYYY HH24:MI:SS'), TO_DATE('03/09/2013 18:33:10', 'MM/DD/YYYY HH24:MI:SS'));
COMMIT;
Query used:
SELECT time_stamp,
       UPPER (SUBSTR(status,1,1)) as status,
       Unique_ID,
       Time_A,
       Time_B,
       CASE
            WHEN time_stamp BETWEEN Time_A AND Time_B
          THEN
               'Y'
          ELSE
               'N'
         END   as flag
  FROM t2;
the query provided above doesn't work with these rows:
TIME_STAMP,                STATUS,     UNIQUE_ID,      TIME_A,                             TIME_B,                                    FLAG
2013/03/09 6:26:55 PM,   L,              6583483,          2013/03/09 6:15:51 PM,     2013/03/09 6:26:28 PM,              N
2013/03/09 6:26:55 PM,   O,              6583483,          2013/03/09 6:27:01 PM,      2013/03/09 6:33:10 PM,             N
because here where time_stamp is same and time_stamp is not between time_a and time_b therefore it should be flagged Y on the row where status is *' O '*
Edited by: 855161 on Apr 10, 2013 7:36 AM

Answer 1

Hi,
Sorry, it's unclear what you want.
855161 wrote:
Hi All
I need some help here:
1. I have to read multiple rows to check where time stamp is same for example rownum (1-3) : *'2013/03/09 10:54:09 PM'*
2. And status is unique for example *(H,L)*
3. And if TIME_STAMP is between TIME_A and TIME_B then flag it as Y else flag it as N.
and if the above condition is not true then
1. Flag the row with ' Y ' where time_stamp is same i.e. : '2013/03/09 10:54:09 PM'
2. And the status is *( ' O ')*
3. the other rows where time_stamp is same, and unique_id is same, and STATUS is either or both ( ' H ' , ' L ' ) flag it as *' N '*
the query provided above doesn't work with these rows:
TIME_STAMP,                STATUS,     UNIQUE_ID,      TIME_A,                             TIME_B,                                    FLAG
2013/03/09 6:26:55 PM,   L,              6583483,          2013/03/09 6:15:51 PM,     2013/03/09 6:26:28 PM,              N
2013/03/09 6:26:55 PM,   O,              6583483,          2013/03/09 6:27:01 PM,      2013/03/09 6:33:10 PM,             Nbecause here where time_stamp is same and time_stamp is not between time_a and time_b therefore it should be flagged Y on the row where status is *' O '*What are the complet, correct results you want from thissample data?
Are they the same as what you are currently getting, except that the 2 rows above should have flag = 'O'?
Why should the row that has status='L' get flag='O'?
Why don't the other rows that currently have status='O' and flag='N' need to be changed?
To see if there are other rows with the same time_stamp, you can use the analytic COUNT function, like this:
SELECT time_stamp,
       UPPER (SUBSTR(status,1,1)) as status,
       Unique_ID,
       Time_A,
       Time_B,
       CASE
            WHEN time_stamp BETWEEN Time_A AND Time_B
                THEN
                    'Y'
            WHEN  COUNT (*) OVER (PARTITION BY  time_stamp) > 1
         AND       UPPER (SUBSTR (status, 1, 1)) = 'O'
             THEN
              'O'
                ELSE
                    'N'
         END   as flag
  FROM t2
ORDER BY  time_stamp
,            status
;The results that this produces are :
`                        UNIQUE
TIME_STAMP           S      _ID TIME_A               TIME_B               F
09-Mar-2013 18:26:55 L  6583483 09-Mar-2013 18:15:51 09-Mar-2013 18:26:28 N
09-Mar-2013 18:26:55 O  6583483 09-Mar-2013 18:27:01 09-Mar-2013 18:33:10 O
09-Mar-2013 22:54:09 H  6587797 09-Mar-2013 23:00:22 09-Mar-2013 23:03:00 N
09-Mar-2013 22:54:09 L  6587797 09-Mar-2013 22:48:35 09-Mar-2013 23:00:22 Y
09-Mar-2013 22:54:09 O  6587797 09-Mar-2013 23:03:00 09-Mar-2013 23:05:54 O
09-Mar-2013 22:54:57 H  6587797 09-Mar-2013 23:00:22 09-Mar-2013 23:03:00 N
09-Mar-2013 22:54:57 L  6587797 09-Mar-2013 22:48:35 09-Mar-2013 23:00:22 Y
09-Mar-2013 22:54:57 O  6587797 09-Mar-2013 23:03:00 09-Mar-2013 23:05:54 OI don't think this is what you wat, but I'm not sure.
Can you explain, in a different way than before, what flag is supposed to show? What role (if any) does the so-called unique_id play in this problem?