Trigger file

Question

Greetings,
I'm currently writing a very simple trigger... I'm using oracle 10.2
In the query analyser, I wrote:
DROP TRIGGER TRIGMAXPARTIE;
CREATE TRIGGER TRIGMAXPARTIE
BEFORE INSERT
ON PARTIE
FOR EACH ROW
        DECLARE nbparties number;
    BEGIN
        select count(NoR) as nbparties
        from PARTIE
        group by NoR;
        IF (nbparties >= 5)
        THEN raise_application_error (-208476, 'Le nombre de parties maximal pouvant être effectué lors d''une rencontre est atteint');
    END IF;
END;but it stops entering the trigger after the "DECLARE nbparties number;"
If I enter it manually, my trigger works (right click, alter trigger)
but I need a single run file...
Any ideas what might cause this?
Thanks!

Answer 1

A database trigger is not the right tool for the job here. See this test case:
SQL> create table partie (NoR) as select level from dual connect by level <= 4
  2  /
Tabel is aangemaakt.
SQL> CREATE TRIGGER TRIGMAXPARTIE
  2     BEFORE INSERT ON PARTIE
  3     FOR EACH ROW
  4  DECLARE
  5     nbparties NUMBER;
  6  BEGIN
  7     SELECT COUNT(NoR) INTO nbparties FROM PARTIE;
  8     IF (nbparties >= 5) THEN
  9        raise_application_error(-20476,
10                                'Le nombre de parties maximal pouvant Ûtre effectuÚ lors d''une rencontre est atteint');
11     END IF;
12  END;
13  /
Trigger is aangemaakt.
SQL> insert into partie values (5)
  2  /
1 rij is aangemaakt.
SQL> insert into partie values (6)
  2  /
insert into partie values (6)
FOUT in regel 1:
.ORA-20476: Le nombre de parties maximal pouvant Ûtre effectuÚ lors d'une rencontre est atteint
ORA-06512: in "RWIJK.TRIGMAXPARTIE", regel 6
ORA-04088: Fout bij uitvoering van trigger 'RWIJK.TRIGMAXPARTIE'.So it looks like the code is working. But watch what happens when you open a second session:
SQL> remark
SQL> remark  Open up a second session and issue:
SQL> remark
SQL> remark  insert into partie values (6);
SQL> remark  commit;
SQL> remark
SQL> pauseOutput of the second session:
SQL> insert into partie values (6);
1 rij is aangemaakt.
SQL> commit;
Commit is voltooid.And the first session again:
SQL> commit;
Commit is voltooid.
SQL> select count(NoR) from partie
  2  /
COUNT(NOR)
         6
1 rij is geselecteerd.I doubt this is what you want.
You may want to go with a on commit fast refreshable materialized view that stores the count of records and place a check constraint on the materialized view. But more important, can you state your business rule. Maybe we can find an even better solution.
Regards,
Rob.