Can this pl/sql code be faster?

From: Paschal Mushubi <Paschal.Mushubi_at_cognos.com>
Date: Wed, 09 Jun 1999 02:50:28 GMT
Message-ID: <375E01BC.D68B77CC_at_cognos.com>

I want to update NEW_TABLE with data from T1 and T2 (aliases) There are over 50,000 records to update. The number may grow.. There are no indexes on the target table except on the primary key. This procedure is too slow. Can you make it faster? Suggestions, please. Here is the example code (the real procedure is much longer but uses same
logic)
DECLARE
CURSOR c_customers is
SELECT rowid customer_rowid,

        customer_number,
        address

FROM NEW_TABLE
v_customers c_customers%ROWTYPE;
BEGIN

        OPEN c_customers;
        FETCH c_customers INTO v_customers;
        OPEN c_customers;
        FETCH c_customers INTO v_customers;
        WHILE c_customers%FOUND LOOP
                BEGIN
                        SELECT  CUST.address
                        INTO    v_customers.address
                        FROM    CUSTOMERS T1,
                                ADDRESES T2
                        WHERE   T1.customer_number =
v_customers.customer_number
                        and     T1.cust_id = T2.cust_id;
                                EXCEPTION
                                WHEN NO_DATA_FOUND THEN
                                   v_customers.address := null;

END;
        BEGIN
             UPDATE  VZN.VZN_GL_TRANSACTIONS
             SET     address = v_customers.address
             WHERE   rowid = v_customers.customer_rowid;
        END;
                v_customers.address := NULL;
        FETCH c_customers INTO v_customers;
        END LOOP;
        CLOSE c_customers;
        COMMIT;

END get_customer_address; Received on Wed Jun 09 1999 - 04:50:28 CEST