New/Old variable in SQL update statement - oracle

Question

How can i update the old row value. I have used the :old variable but it does not work for me. Herein my code, if PID already exists in project table i.e. (varProjectExists = 1) then i update the payment term. Now, if payment term is null , i want to update with previous paymnent term value else new payment term that is flown
CREATE OR REPLACE TRIGGER TRIG_PROJECT_INSERT AFTER
INSERT ON TEST_SYN_EAI_PROJECT_IN FOR EACH row DECLARE varError_Msg NVARCHAR2(100);
varSucceeded NVARCHAR2(1);
varActive_YN NVARCHAR2(50);
varProject_Id INT;
varPid INT ;
varPay_Term VARCHAR2(200);
varPay_Term1 VARCHAR2(200);
varError_id INT;
varCurr_activeyn INT;
varProjectExists NUMBER;
BEGIN
varError_Msg := 'No error';
varSucceeded := 'Y';
varError_id := 0;
varProjectExists := 0;
varPID := :new.pid;
varPay_Term := :new.ATTRIBUTE1;
varPay_Term1 := :old.ATTRIBUTE1;
varActive_YN := :new.active_yn;
varProject_ID := :new.project_id;
IF (NVL(varProject_Id,0) = 0 ) THEN
varError_Msg := 'project ID can not be null';
varSucceeded := 'N';
varError_id := 1;
END IF;
SELECT
CASE
WHEN (UPPER(varActive_YN) = 'ACTIVE'
OR UPPER(varActive_YN) = 'Y')
THEN 1
WHEN (UPPER(varActive_YN) = 'INACTIVE'
OR UPPER(varActive_YN) = 'N')
THEN 0
ELSE varcurr_activeyn
END
INTO varActive_YN
FROM Dual;
SELECT COUNT(1)
INTO varProjectExists
FROM project
WHERE ProjectUniversalID = varProject_ID;
IF (varProjectExists = 1) THEN
UPDATE project
SET PID = varPID,
PAYMENTTERM =
CASE
WHEN varPay_Term = 'NULL'
THEN varPay_Term1
WHEN varPay_Term IS NULL
THEN varPay_Term1
ELSE varPay_Term
END
ELSE .....

Answer 1

The OLD pseudo-record refers to the pre-update state of the current row, and so is only meaningful during an update or delete operation. If you're inserting a new record there is no 'old' state to refer to.
From your comment you want to use the previous payment term value that already exists in the PROJECT table. That would't be available in OLD anyway as it isn't the table the trigger is against. If that is the case then you need to retrieve it at the same time you check that it exists. Something like:
CREATE OR REPLACE TRIGGER TRIG_PROJECT_INSERT
AFTER INSERT ON TEST_SYN_EAI_PROJECT_IN
FOR EACH ROW
DECLARE
...
BEGIN
...
varPay_Term := :new.ATTRIBUTE1;
-- varPay_Term1 := :old.ATTRIBUTE1; -- not valid
varActive_YN := :new.active_yn;
...
-- this doesn't need to select from dual, you can assign directly
varCurr_activeyn := CASE
WHEN (UPPER(varActive_YN) = 'ACTIVE'
OR UPPER(varActive_YN) = 'Y')
THEN 1
WHEN (UPPER(varActive_YN) = 'INACTIVE'
OR UPPER(varActive_YN) = 'N')
THEN 0
ELSE varCurr_activeyn
END CASE;
-- get the current term as you check if a record exists
SELECT COUNT(1), MAX(PAYMENTTERM)
INTO varProjectExists, varPay_Term1
FROM project
WHERE ProjectUniversalID = varProject_ID;
IF (varProjectExists = 1) THEN
UPDATE project
SET PID = varPID,
PAYMENTTERM =
CASE
WHEN varPay_Term = 'NULL'
THEN varPay_Term1
WHEN varPay_Term IS NULL
THEN varPay_Term1
ELSE varPay_Term
END
...
The MAX() is needed because you're already using an aggregate COUNT(), to avoid a no-data-found error if the ID doesn't exist. Assuming that ID is unique, it doesn't matter if you use MAX or MIN, the result is the same. (If it isn't unique then you'd have to decide which term value to use, so it's probably a safe assumption).