Triggers in sql: Mutating Trigger issue - oracle

Question

I want to insert rows into another table called excercise_scores when a row is inserted in attempts table. There are multiple tries for an excercise and based on the scoring method(which can be latest score, average of all attempts or maximum score from all attempts), I have to insert score in the excercise_scores table.
So when Inserting scores into excercise_score, I also have to check what is the grading method for the excercise, get the score from attempts table based on the method and then finally insert into excercise_scores table.
For that I have created a SQL trigger as follows
CREATE OR REPLACE TRIGGER scores_trigger AFTER
INSERT
ON attempts FOR EACH row DECLARE P1 VARCHAR2(50);
SCORES FLOAT;
v_exists VARCHAR2(1) := 'F';
BEGIN
BEGIN
SELECT
'T'
INTO
v_exists
FROM
excercise_scores
WHERE
user_id = :NEW.USER_ID
AND EXCERCISE_ID = :NEW.EXCERCISE_ID
AND COURSE_ID = :NEW.COURSE_ID;
EXCEPTION
WHEN no_data_found THEN
NULL;
END;
SELECT
SELECTION_METHOD
INTO
P1
FROM
EXCERCISES
WHERE
EXCERCISE_ID = :NEW.EXCERCISE_ID
AND COURSE_ID = :NEW.COURSE_ID;
IF (P1 = 'Latest') THEN
SCORES := :NEW.TOTAL_POINTS;
ELSE
IF (P1 = 'best') THEN
SELECT
MAX(TOTAL_POINTS)
INTO
SCORES
FROM
ATTEMPTS
WHERE
EXCERCISE_ID = :NEW.EXCERCISE_ID
AND COURSE_ID = :NEW.COURSE_ID
AND USER_ID = :NEW.USER_ID;
ELSE
SELECT
AVG(TOTAL_POINTS)
INTO
SCORES
FROM
ATTEMPTS
WHERE
EXCERCISE_ID = :NEW.EXCERCISE_ID
AND COURSE_ID = :NEW.COURSE_ID
AND USER_ID = :NEW.USER_ID;
END IF;
END IF;
IF v_exists = 'T' THEN
UPDATE
EXCERCISE_SCORES
SET
TOTAL_POINTS = SCORES
WHERE
user_id = :NEW.USER_ID
AND EXCERCISE_ID = :NEW.EXCERCISE_ID
AND COURSE_ID = :NEW.COURSE_ID;
ELSE
INSERT
INTO
EXCERCISE_SCORES VALUES
(
:NEW.EXCERCISE_ID,
:NEW.COURSE_ID,
:NEW.COURSE_ID,
SCORES
) ;
END IF;
END;
But when I try to execute this trigger, it gives me the following error:
Error starting at line 47 in command:
insert into attempts values(21,0,'vshesha',3,'CSC540',TO_DATE('20141014', 'YYYYMMDD'),12,6)
Error report:
SQL Error: ORA-04091: table VSHESHA.ATTEMPTS is mutating, trigger/function may not see it
ORA-06512: at "VSHESHA.SCORES_TRIGGER", line 18
ORA-04088: error during execution of trigger 'VSHESHA.SCORES_TRIGGER'
04091. 00000 - "table %s.%s is mutating, trigger/function may not see it"
*Cause: A trigger (or a user defined plsql function that is referenced in
this statement) attempted to look at (or modify) a table that was
in the middle of being modified by the statement which fired it.
*Action: Rewrite the trigger (or function) so it does not read that table.
I couldn't figure out why it is giving this error as I am running the trigger after insert. Generally mutating table issue should not come in after insert triggers.
Can anyone please help me with this. I am suck with this for a long time now.

Answer 1

The error message is pretty self-explaining, and the SQL statement that's causing it is:
SELECT
AVG(TOTAL_POINTS)
INTO
SCORES
FROM
ATTEMPTS
WHERE
EXCERCISE_ID = :NEW.EXCERCISE_ID
AND COURSE_ID = :NEW.COURSE_ID
AND USER_ID = :NEW.USER_ID;
You're selecting from ATTEMPTS, and that's the table that's currently changing. There's no simple solution for it - the most sensible approach is to use move your business logic out of the trigger and into your application code (see AskTom about mutating table errors ).
BTW: Your statement that "Generally mutating table issue should not come in after insert triggers." is plain wrong. See database journal article for a longer explanation (in short: you will always run into mutating table errors with multi-row inserts, and in almost every case run into errors with single-row inserts).

Answer 2

In Oralce you can use COMPOUND TRIGGER. In your case it would be similar to this one. I did not copy all your operations to this example, but you should get the principle.
However, as the other replies already mentionied. The preferred way should be to move all this logic into a PL/SQL procedure.
CREATE OR REPLACE TRIGGER scores_trigger
FOR INSERT ON attempts
COMPOUND TRIGGER
TYPE RowTableType IS TABLE OF ROWID;
RowTable RowTableType;
aRow attempts%ROWTYPE;
SCORES FLOAT;
v_exists VARCHAR2(1) := 'F';
BEFORE STATEMENT IS
BEGIN
RowTable := RowTableType();
END BEFORE STATEMENT;
AFTER EACH ROW IS
BEGIN
RowTable.EXTEND;
RowTable(RowTable.LAST) := :NEW.ROWID;
END AFTER EACH ROW;
AFTER STATEMENT IS
BEGIN
FOR i IN RowTable.FIRST..RowTable.LAST LOOP
SELECT *
INTO aRow
FROM attempts
WHERE ROWID = RowTable(i);
SELECT SELECTION_METHOD
INTO P1
FROM EXCERCISES
WHERE EXCERCISE_ID = aRow.EXCERCISE_ID
AND COURSE_ID = aRow.COURSE_ID;
IF P1 = 'Latest' THEN
SCORES := :NEW.TOTAL_POINTS;
ELSE
IF (P1 = 'best') THEN
SELECT MAX(TOTAL_POINTS)
INTO SCORES
FROM ATTEMPTS
WHERE EXCERCISE_ID = aRow.EXCERCISE_ID
AND COURSE_ID = aRow.COURSE_ID
AND USER_ID = aRow.USER_ID;
ELSE
SELECT AVG(TOTAL_POINTS)
INTO SCORES
FROM ATTEMPTS
WHERE EXCERCISE_ID = aRow.EXCERCISE_ID
AND COURSE_ID = aRow.COURSE_ID
AND USER_ID = aRow.USER_ID;
END IF;
END IF;
END LOOP;
END AFTER STATEMENT;
END scores_trigger;
An example for an PL/SQL procedure would be this one:
CREATE OR REPLACE PROCEDURE insert_score(V_EXCERCISE IN SCORES.EXCERCISE_ID%TYPE, V_COURSE IN SCORES.COURSE_ID%TYPE, V_USER IN USER_ID%TYPE) IS
BEGIN
INSERT INTO insert_score (EXCERCISE_ID, COURSE_ID, USER_ID) VALUES (V_EXCERCISE, V_COURSE, V_USER);
BEGIN
SELECT 'T'
INTO v_exists
FROM excercise_scores
WHERE user_id = :NEW.USER_ID
AND EXCERCISE_ID = :NEW.EXCERCISE_ID
AND COURSE_ID = :NEW.COURSE_ID;
EXCEPTION
WHEN NO_DATA_FOUND THEN
RETURN;
END;
... other stuff
END insert_score;
For the insert you have to call this procedure instead of doing the insert directly.