How to calculate opening and closing balance for period

Question

Hi all,
i have to find out opening and closing balance.
the table structure of temporary table is
select * from hwcn_xn_fa_report_temp1 where asset_id=10029400
PERIOD_COUNTER CST_OP_BAL CST_ADDITION CST_TRANSFER CST_DISPOSAL COST_CLOSING_BALANCE
24108 0 0 0
24109 12000
24110 0 0 0
24111 0 0 0
in this table cst_op_balnce and cost_closing_balace is null
i have display cost_op_bal and cost_closing_balnce
cost_closing_balance=cst_op_bal+cst_addition+cst_transfer+cst_disposal
for period 2408 op_balnce=0 closing_bal=0
for period 2409 op_balnce=0 closing_balce=1200
for period 2410 op_bal=1200 closing_bal=1200
closing balance of dec will be opening bal of jan
thanks and regards
Edited by: user10664276 on Apr 19, 2009 11:08 PM
Edited by: user10664276 on Apr 19, 2009 11:13 PM

Answer 1

Hi,
user11118871 wrote:
Can you explain what that is? Thank you if you have one example.
ROWS BETWEEN  UNBOUNDED PRECEDING AND 1 PRECEDING
When you use the analytic SUM (c) function, then, on each row, it returns the values of column (or expression) c from several rows in the result set added together.
Which rows? That depends.
If the analytic clause (the part in parentheses after OVER) does not include ORDER BY, then it is all rows.
If the analytic clause has an ORDER BY clause, but no windowing clause (that is, ROWS BETWEEN ... or RANGE BETWEEN ...), then the rows included in the sum are all rows up to and including the row where the function is being called (as sorted by the analytic ORDER BY).
If the analytic cluase has both ORDER BY and a windowing clause "ROWS BETWEEN x PRECEDING AND y PRECEDING", then the rows included in the sum are the rows from x to y rows before the one where the function is called.
Do some experiments with different values of x and y.
First, create a table like the one in the problem above, but simplified a little.
CREATE TABLE     test_sum
(      period     NUMBER
,      new_amt     NUMBER
INSERT INTO test_sum (period, new_amt) VALUES (24108,     1);
INSERT INTO test_sum (period, new_amt) VALUES (24109,     4);
INSERT INTO test_sum (period, new_amt) VALUES (24110,     2);
INSERT INTO test_sum (period, new_amt) VALUES (24111,     8);
INSERT INTO test_sum (period, new_amt) VALUES (25001,     32);
INSERT INTO test_sum (period, new_amt) VALUES (25002,     16);
COMMIT;The original problem above used names that were meaningful for its application, and columns that have nothing to do with the SUM function. Let's simplify the former and lose the latter.
That problem involved the SUM of three columns added together. Since we just want to understand how the windowing clause works, let's simplify that to one column.
With these simplifications, my original query is:
SELECT       period
,       new_amt     
,       SUM (new_amt) OVER ( ORDER BY          period
                                     ROWS BETWEEN  UNBOUNDED PRECEDING
                             AND          1            PRECEDING
                    ) AS opening_balance
,       SUM (new_amt) OVER ( ORDER BY          period
                    ) AS closing_balance
FROM       test_sum
ORDER BY  period;Given the data above, it produces these results:
.   PERIOD    NEW_AMT OPENING_BALANCE CLOSING_BALANCE
     24108          1                               1
     24109          4               1               5
     24110          2               5               7
     24111          8               7              15
     25001         32              15              47
     25002         16              47              63So, for example, on the row where period=24110,
opening_balance=5, which is the total of new_amt from all rows up to but not including that row: 5=1+4, and
closing_balance=7, which is the total of new_amt from all rows up to and including that row: 7=1+4+2.
To really understand how the windowing clause works, do some experiments. Change the definition of opening_balance to include " BETWEEN x PRECEDING AND y PRECEDING". You'll find that:
(a) "UNBOUNDED PRECEDING" means the same as "n PRECEDING", where n is greater than the number of rows in your result set.
(b) "CURRENT ROW" means the same as "0 PRECEDING"
(c) x must be greater than or equal to y
(d) neither x nor y can be negative (but you can use "FOLLOWING" instead of "PRECEDING" to get the same effect).
For more, see the introduction to "Analytic Functions" in the [SQL Language manual|http://download.oracle.com/docs/cd/B28359_01/server.111/b28286/functions001.htm#sthref972]
When you're finished, don't forget to
DROP TABLE     test_sum;