Two thinks

Question

Hi, I have two questions...
First,
Is Correct to use CASE WHEN in statement Update before SET?, that is:
UPDATE MyTable CASE
WHEN v_month = 0 THEN SET Charges_0 = amount
WHEN v_month = 1 THEN SET Charges_1 = amount
WHEN v_month = 2 THEN SET Charges_2 = amount
WHEN v_month = 3 THEN SET Charges_3 = amount
WHEN v_month = 4 THEN SET Charges_4 = amount
WHEN v_month = 5 THEN SET Charges_5 = amount
WHEN v_month = 6 THEN SET Charges_6 = amount
WHEN v_month = 7 THEN SET Charges_7 = amount
WHEN v_month = 8 THEN SET Charges_8 = amount
WHEN v_month = 9 THEN SET Charges_9 = amount
WHEN v_month = 10 THEN SET Charges_10 = amount
WHEN v_month = 11 THEN SET Charges_11 = amount
ELSE
SET Charges_12 = amount
END
WHERE ID_PERIOD = :v_1 AND ID = :V_2;
if that not is possible in Oracle, what should may be to do?,
To do Dinamic (EXECUTE IMMEDIATE) and build sql statement in a loop or to do hard and build every update's statement.
Or any idea for this, I apreciated
This code may be executed 225000 per year. (18750 every month)...
And Second,
I create a view and have this:
My_Ctx.put_VID('1');
SELECT id, nombre, code FROM vw_MyView FOR UPDATE NOWAIT;
DELETE FROM vw_MyView WHERE id = 1;
--MyCtx, is My Context's, and put_xxxx are method's for input data.
--and get_xxxx are method's for get data.
CREATE VIEW MyView
IS
SELECT
T1.ID ID,
T2.name Name,
T3.Code Mark,
T2.AnyDataMore InitData,
T3.AnyDateMore InitDate
FROM
T_MyTable1 T1,
T_MyTable2 T2,
T_MyTable3 T3
WHERE
T2.ID = T1.ID
AND T3.ID = T2.ID
AND T1.ID = MyCtx.get_vID();
For this example, use the context is for practices, because in a production, the view needed more "parameters", Initial date, Final date, etc, etc... but i don't have this in production, I Have delete directly from the table, obviously SELECT...FOR UPDATE before....
How known Oracle's Server what record delete in what table?
OS: Windows XP
ORACLE SERVER: 9i

Answer 1

Thanks David... it's Great idea....
Only one question, but first an introduction... I'm sorry my enlgish, ok!!...
My database have 3 tables
The Table T_Movs have:
Year
2000 205139 Records
2001 204257 Records
2002 221180 Records
2003 213617 Records
2004 113799 Records and more;
The Table T_Pol have:
2000 26169 Records
2001 24894 Records
2002 23040 Records
2003 20031 Records
2004 10345 Records and more;
The Table T_CC have:
2000 27808 Records
2001 39404 Records
2002 39527 Records
2003 19638 Records
2004 20079 Records and more;
T_Movs REFERENCES TO T_Pol
T_CC REFERENCES TO T_MOVS
For specific cases, I Need Recalculate all year (one year, no all years)
And Then I'll may be this
CURSOR cur_Pol(aYear VARCHAR2(4)) IS
SELECT *
FROM T_POL
WHERE id_Period = aYear;
CURSOR cur_Movs(aYear VARCHAR2(4), aID NUMBER(6)) IS
SELECT *
FROM T_MOVS
WHERE id_period = aYear
AND id_Pol = aID;
Period := '2004';
OPEN cur_Pol(Period);
LOOP
FETcH cur_Pol INTO v_Pol
EXIT WHEN cur_Pol%NOTFOUND
OPEN cur_Movs(Period, v_Pol.ID);
v_month := TO_NUMBER(TO_CHAR(v_Pol.DateAplication,'MM'));
LOOP
FETcH cur_Movs INTO v_Mov
EXIT WHEN cur_Pol%NOTFOUND
IF v_Mov.Code = 0 THEN
UPDATE
T_CC
Set
Charges_0 = Decode(v_month,0,v_Mov.amount,Charges_0),
Charges_1 = Decode(v_month,1,v_Mov.amount,Charges_1),
WHERE ID_PERIOD = Period AND ID = v_Mov.ID_CC;
ELSE
UPDATE
MyTable
Set
Debits_0 = Decode(v_month,0,v_Mov.amount,Debits_0),
Debits_1 = Decode(v_month,1,v_Mov.amount,Debits_1),
WHERE ID_PERIOD = Period AND ID = v_Mov.ID_CC;
END IF;
LOOP;
CLOSE cur_Movs;
END LOOP;
CLOSE cur_Pol;
The idea is,
For Each Pol In the year (in this example, # recors of Pol for '2004' are 10345)
For Each Movs in Pol (in this example #Movs for '2004' are 113749)
then Update
Ok, my example is ilustrated the number of iterations, and for each record has 13 columns for update insted of one.
What happend with performance?... and then is correct the use DECODE's function in this case?...
Thanks
Daniel
PD: Hope I can explain myself