Can it be achieved

Question

Hi Gurus.
CREATE TABLE LOAN_TXN
TXN_DATE     DATE,
BALANCE          NUMBER(10,2),
CODE          VARCHAR2(1),
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '02/15/2010 11:00:00 AM', 'MM/DD/YYYY HH:MI:SS AM'), 250000, 'D');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '03/31/2010 11:59:59 AM', 'MM/DD/YYYY HH:MI:SS AM'), 250000, 'B');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '05/14/2010 11:25:00 AM', 'MM/DD/YYYY HH:MI:SS AM'), 500000, 'D');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '06/30/2010 12:15:00 PM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '08/02/2010 10:45:26 AM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '09/08/2010 02:10:17 PM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '10/27/2010 04:25:20 PM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '11/09/2010 10:15:55 AM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '12/29/2010 03:10:20 PM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '01/12/2011 01:11:15 PM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '02/11/2011 12:11:48 PM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '03/03/2011 12:25:36 PM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '03/31/2011 11:00:00 AM', 'MM/DD/YYYY HH:MI:SS AM'), 4000, 'R');
INSERT INTO LOAN_TXN ( TXN_DATE, BALANCE, CODE) VALUES (TO_Date( '03/31/2011 11:59:59 AM', 'MM/DD/YYYY HH:MI:SS AM'), 460000, 'B');Using the above details, I have been asked to work out a single SQL Query which could generate the Details in the following format.
TXN_DATE     ADVANCE          RECOV     BALANCE          DAYS     AMT     CUMM_AMT
15/02/2010     250000               250000          45     3082     
31/03/2010                    250000               3082               
01/04/2010                    250000          43     2945     3082
14/05/2010     250000               500000          47     6438     6027
30/06/2010               4000     496000          33     4484     12465
02/08/2010               4000     492000          37     4987     16949
08/09/2010               4000     488000          49     6551     28487
27/10/2010               4000     484000          13     1724     30211
09/11/2010               4000     480000          50     6575     36786
29/12/2010               4000     476000          14     1826     38612
12/01/2011               4000     472000          30     3879     42491
11/02/2011               4000     468000          20     2564     45055
03/03/2011               4000     464000          28     3559     48614
31/03/2011               4000     460000          1     126     48740     
31/03/2011                    460000               45658
01/04/2011                    460000          94     11847     Legend : Code 'R' is for Recovery
     Code 'D' is for Disbursement
     Code 'B' is for Balance
AMT is to be calculated as follows :
Difference of Current with the Next Leading Date multiplied with Balance on the date * Interest Rate (which is 10% here) / Days in the Year.
On the last day of the Financial Year i.e. 31/03/YYYY, Closing Balance is to computed summing
the date-wise Amt during the Financial Year i.e. 1st April to 31st Mar.
Can the above be achieved ?

Answer 1

Hi,
Thanks for posting the sample data in such a useful format; that's very helpful!
It would also be helpful if you explained how you get the output all the columns, not just amt.
Like Justin, I don't understand how you get days.
I also don't understand the extra rows you want to add for April 1 after each financial year:
user3308033 wrote:
TXN_DATE     ADVANCE          RECOV     BALANCE          DAYS     AMT     CUMM_AMT
01/04/2010                    250000          43     2945     3082
01/04/2011                    460000          94     11847     
Why do you want cumm_amt for 2010, but not for 2011? Also, as Dan said, how do you get days=94 on the last row?
If there happend to be a transaction on April 1, what would the output look like? Would there be an extra row plus the row(s) that had the April 1 txn_date? Include an example in your sample data and desired results.
This is a step in the right direction:
WITH     got_fiscal_year          AS
     SELECT     txn_date, balance, code
     ,     TRUNC ( ADD_MONTHS (txn_date, +9)
                , 'YEAR'
                )          AS fiscal_year
     FROM     loan_txn
,     got_balance          AS
     SELECT       txn_date, fiscal_year
     ,       CASE
                WHEN  code = 'D'
                           THEN  balance
            END                              AS advance
     ,       CASE
                WHEN  code = 'R'
                THEN  balance
              END                              AS recov
     ,       SUM ( CASE  code
                         WHEN  'D'  THEN   balance
                       WHEN  'R'  THEN  -balance
                     END
                )     OVER ( PARTITION BY  fiscal_year
                                   ORDER BY          txn_date
                         )                         AS balance
     ,       CEIL  ( LEAD (txn_date) OVER ( PARTITION BY  fiscal_year
                                               ORDER BY      txn_date
               - txn_date
               )                         AS days
     ,       ADD_MONTHS (fiscal_year, 12) - fiscal_year     AS days_in_year
     FROM       got_fiscal_year
,     got_amt          AS
     SELECT       got_balance.*
     ,       ( .10     -- 10% interest rate
            * balance
            * days
            / days_in_year
                  )                     AS amt
     FROM       got_balance
SELECT       txn_date
,       advance
,       recov
,       balance
,       days
,       amt
,       SUM (amt) OVER ( PARTITION BY  fiscal_year
                            ORDER BY        txn_date
                )       AS cumm_amt
FROM      got_amt
ORDER BY  txn_date
;Output:
`                                                CUMM
                                                 _AMT
TXN_DATE    ADVANCE RECOV BALANCE DAYS    AMT
15-Feb-2010  250000        250000   45   3082    3082
31-Mar-2010                250000                3082
14-May-2010  500000        500000   48   6575    6575
30-Jun-2010          4000  496000   33   4484   11060
02-Aug-2010          4000  492000   38   5122   16182
08-Sep-2010          4000  488000   50   6685   22867
27-Oct-2010          4000  484000   13   1724   24591
09-Nov-2010          4000  480000   51   6707   31298
29-Dec-2010          4000  476000   14   1826   33123
12-Jan-2011          4000  472000   30   3879   37003
11-Feb-2011          4000  468000   21   2693   39695
03-Mar-2011          4000  464000   28   3559   43255
31-Mar-2011          4000  460000    1    126   43381
31-Mar-2011                460000               43381As you can see, I'm not getting days exactly like you want, so the amts are off accordingly. Also, the best way to add the extra rows for April 1 depend on exactly what those rows are supposed to show. Since I don't understand that, I didn't include the April 1 rows at all.
We might not need this many sub-queries. I'll wait until I understand the problem before trying to simplify it.