foreach rows of my table and update my ROW_NUMBER column - sql

Question

I would like to create a script pl/sql where I can modify the value of my column ROW_NUMBER (the first time the value of ROW_NUMBER equal NULL).
This is the structure of my table 'A' :
CREATE TABLE A
(
"NAME" VARCHAR2(25 BYTE),
"NUM" NUMBER(10,0)
)
I would like to foreach all rows of table A and increment my Column 'NUM' by 1 if Column 'NAME' equal 'DEB'.
I would like to get the result like :
I created one pl/sql script :
DECLARE
INcrmt NUMBER(4):=1;
line WORK_ODI.TEST_SEQ%ROWTYPE;--before fetch it returns 0
CURSOR c_select IS
SELECT ROW_NUMBER,VALUE FROM WORK_ODI.TEST_SEQ;
BEGIN
OPEN c_select;
LOOP
FETCH c_select INTO line;
DBMS_OUTPUT.PUT_LINE(line.VALUE);
if line.VALUE like '%DEB%'
then
UPDATE WORK_ODI.TEST_SEQ SET ROW_NUMBER = INcrmt WHERE VALUE=line.VALUE;
INcrmt := INcrmt + 1;
end if;
if line.VALUE not like '%DEB%'
then
UPDATE WORK_ODI.TEST_SEQ SET ROW_NUMBER = INcrmt WHERE VALUE=line.VALUE;
end if;
EXIT WHEN c_select%NOTFOUND;
END LOOP;
CLOSE c_select;
COMMIT;
END;
DECLARE
INcrmt NUMBER(4):=1;
line WORK_ODI.TEST_SEQ%ROWTYPE;--before fetch it returns 0
CURSOR c_select IS
SELECT ROW_NUMBER,VALUE FROM WORK_ODI.TEST_SEQ;
BEGIN
OPEN c_select;
LOOP
FETCH c_select INTO line;
DBMS_OUTPUT.PUT_LINE(line.VALUE);
if line.VALUE like '%DEB%'
then
UPDATE WORK_ODI.TEST_SEQ SET ROW_NUMBER = INcrmt WHERE VALUE=line.VALUE;
INcrmt := INcrmt + 1;
end if;
if line.VALUE not like '%DEB%'
then
UPDATE WORK_ODI.TEST_SEQ SET ROW_NUMBER = INcrmt WHERE VALUE=line.VALUE;
end if;
EXIT WHEN c_select%NOTFOUND;
END LOOP;
CLOSE c_select;
COMMIT;
END;
but this is not work well , please take a look at what it gives me as result :
please anybody can help me

Answer 1

First, you should have an Aid column of some sort. In Oracle 12+, you can use an identity. In earlier versions, you can use a sequence. This provides an ordering for the rows in the table, based on insert order.
Second, you can do what you want on output:
select a.*,
sum(case when a.name like 'DEB%' then 1 else 0 end) over (order by aid) as row_number
from a;
If you really need to keep the values in the table, then you can use a merge statement to assign values to existing rows (the aid column is very handy for this). You will need a trigger afterwards to maintain it.
My suggestion is to do the calculation on the data, rather than storing the value in the data. Maintaining the values with updates and deletes seems like a real pain.