How to insert an ID into another table - sql

Question

I'm practising triggers on Oracle and I'm having troubles with this trigger:
CREATE OR REPLACE TRIGGER MODIFICACIONES_SALARIOS AFTER
UPDATE OF salario ON empleados_pac
FOR EACH ROW
DECLARE
v_username VARCHAR(10);
v_hora VARCHAR2(10);
BEGIN
SELECT user INTO v_username FROM dual;
SELECT to_char(sysdate, 'HH24:MI:ss') INTO v_hora FROM dual;
INSERT INTO audita_salarios (id_emp, salario_antiguo, salario_nuevo, fecha, hora, username) VALUES (id_empleado, :old.salario, :new.salario, sysdate, v_hora, v_username);
END;
I'm doing an historic table where when the salary of the employeer changes the trigger inserts in the historic his ID, the old salary, the new one, the date and the user who changed the salary.
All works perfectly except for the ID insert. Oracle says that the column is not allowed.
I think the problem is because I want to insert an unique ID into a new column where the ID is going to be repeated if the salary changes two or more times. How could I fix this?

Answer 1

I think the problem is because I want to insert an unique ID into a new column where the ID is going to be repeated if the salary changes two or more times. How could I fix this?
Do not have a UNIQUE constraint on audita_salarios.id_emp; instead have a composite UNIQUE constraint on id_emp and fecha (if you need a UNIQUE constraint at all).
Also:
use a virtual column for HORA (or remove that column as it is just duplicating the time component of FECHA);
use :NEW.id_empleado rather than id_empleado in the trigger.
CREATE TABLE empleados_pac (
id_empleado NUMBER(2),
salario NUMBER(10,2)
);
CREATE TABLE audita_salarios (
id_emp NUMBER(2),
salario_antiguo NUMBER(10,2),
salario_nuevo NUMBER(10,2),
fecha DATE,
hora VARCHAR2(8)
GENERATED ALWAYS AS ( CAST( TO_CHAR( fecha, 'HH24:MI:SS' ) AS VARCHAR2(8) ) ),
username VARCHAR2(60),
CONSTRAINT audita_salarios__pk PRIMARY KEY ( id_emp, fecha )
);
CREATE TRIGGER MODIFICACIONES_SALARIOS
AFTER UPDATE OF salario ON empleados_pac
FOR EACH ROW
DECLARE
BEGIN
INSERT INTO audita_salarios (
id_emp, salario_antiguo, salario_nuevo, fecha, username
) VALUES (
:new.id_empleado, :old.salario, :new.salario, sysdate, USER
);
END;
/
INSERT INTO empleados_pac ( id_empleado, salario ) VALUES ( 1, 10 );
UPDATE empleados_pac
SET salario = 20
WHERE id_empleado = 1;
BEGIN
DBMS_SESSION.SLEEP(1);
END;
/
UPDATE empleados_pac
SET salario = 30
WHERE id_empleado = 1;
-- You do not actually need to constraint on the audit table unless you
-- want to enforce that salary updates must be on different times.
ALTER TABLE audita_salarios DROP CONSTRAINT audita_salarios__pk;
UPDATE empleados_pac
SET salario = 40
WHERE id_empleado = 1;
Then:
SELECT * FROM audita_salarios;
Outputs:
ID_EMP | SALARIO_ANTIGUO | SALARIO_NUEVO | FECHA | HORA | USERNAME
-----: | --------------: | ------------: | :------------------ | :------- | :--------------------------
1 | 10 | 20 | 2021-04-01 13:02:42 | 13:02:42 | FIDDLE_XRWUFWRAYDJNIGNCOQFP
1 | 20 | 30 | 2021-04-01 13:02:43 | 13:02:43 | FIDDLE_XRWUFWRAYDJNIGNCOQFP
1 | 30 | 40 | 2021-04-01 13:02:43 | 13:02:43 | FIDDLE_XRWUFWRAYDJNIGNCOQFP
db<>fiddle here