When is the SQL cursor loads the data? - oracle11g

Question

I have following script to update local table, by reading a remote source table. Script is running fine without any error. This post is to clarify properly about, How the cursor works.
I read data from remote source table Source_Product and initially insert into temp table VW_PRODUCT. After that I insert or update my PRODUCT table.
My questions are.
1) When does the product_cursor load the data? Is it when we try to read the cursor in the for loop? or when ever declare the cursor it loads the data?
2) This script runs daily. If the product_cursor runs as soon as it declare, then VW_PRODUCT has previous day data. Because still the today data is not inserted to VW_PRODUCT table ( insert query is available after the cursor declaration). So the product_cursor will not have any record after minus. Because ysterday_data minus ysterday_data is zero. So how can it update or insert the latest data to PRODUCT according to below script?
SET serveroutput ON SIZE 1000000;
DECLARE
CURSOR product_cursor
IS
SELECT V.PRODUCTID,
V.PACKAGEID'
V.ENDDATE
FROM VW_PRODUCT V
MINUS
SELECT E.PRODUCTID,
E.PACKAGEID,
E.ENDDATE
FROM PRODUCT E;
/*The delete data*/
CURSOR product_cursor_del
IS
SELECT E.PRODUCTID FROM PRODUCT E WHERE (E.ENDDATE > SYSDATE OR E.ENDDATE IS NULL)
MINUS
SELECT V.PRODUCTID FROM VW_PRODUCT V;
/* Variable Declaration*/
v_total NUMBER (10);
v_inserted NUMBER (10);
v_updated NUMBER (10);
v_deleted NUMBER (10);
v_rows_inserted NUMBER (10);
v_productid PRODUCT.PRODUCTID%TYPE;
v_count NUMBER (10);
v_commit_point NUMBER := 25;
BEGIN
v_total := 0;
v_count := 0;
v_inserted := 0;
v_updated := 0;
v_deleted := 0;
v_rows_inserted := 0;
EXECUTE IMMEDIATE 'TRUNCATE TABLE VW_PRODUCT';
INSERT INTO VW_PRODUCT
SELECT * FROM Source_Product;
SELECT COUNT (*)
INTO v_rows_inserted
FROM VW_PRODUCT;
COMMIT;
/*delete data*/
FOR product_rec IN product_cursor_del
LOOP
BEGIN
v_total := v_total + 1;
update product set enddate = sysdate
WHERE productid = product_rec.productid and enddate is null;
v_deleted := v_deleted + 1;
v_count := v_count + 1;
IF (v_count >= v_commit_point)
THEN
COMMIT;
v_count := 0;
END IF;
EXCEPTION
WHEN OTHERS
THEN
BEGIN
DBMS_OUTPUT.put_line ( 'Exception with product: ' );
END;
END;
END LOOP;
FOR product_rec IN product_cursor
LOOP
BEGIN
v_total := v_total + 1;
SELECT productid
INTO v_productid
FROM product
WHERE productid = product_rec.productid;
update PRODUCT
set PACKAGEID = product_rec.PACKAGEID,
ENDDATE = product_rec.ENDDATE
WHERE PRODUCTID = product_rec.PRODUCTID;
v_updated := v_updated + 1;
EXCEPTION
WHEN NO_DATA_FOUND
THEN
INSERT INTO PRODUCT
(PRODUCTID,PACKAGEID,ENDDATE)
VALUES
(product_rec.PRODUCTID,
product_rec.PACKAGEID,
product_rec.ENDDATE);
v_inserted := v_inserted + 1;
v_count := v_count + 1;
IF (v_count >= v_commit_point)
THEN
COMMIT;
v_count := 0;
END IF;
WHEN OTHERS
THEN
raise_application_error ('Error );
END;
END LOOP;
IF (v_total >= 1)
THEN
COMMIT;
END IF;
END;
/

Answer 1

In a nutshell, to answer your basic questions,
A cursor doesn't store any result set, it is a pointer used to fetch rows from a result set.
Memory is not consumed at declare stage.
It is the FETCH statement when you are actually using the cursor. The FETCH statement retrieves rows from the result set and places them into an area in the memory. You could fetch the rows one at a time, several at a time, or all at once.
The FETCH statement performs the following operations:
Reads the data for the current row in the result set into the output
PL/SQL variables.
Moves the pointer to the next row in the result set.
Take a look at working with cursors.