> This means upd1 is redundant since any rows affected by upd1 *must* be > affected by upd2. OK
> > CREATE RULE transactions_sco_up8 AS ON > > UPDATE TO transactions_sco_v DO INSTEAD UPDATE transactions_sco SET traiter > > = new.traiter WHERE > > (transactions_sco.id = old.id); > > OK, this one just compares "id", which is presumably the primary key and > unique. Right "id" is the primary key > > Now look what is happening: > > > > SELECT count(1) from transactions_sco where traiter='f'; > > count > > ------- > > 17591 > > > > update transactions_sco_v set traiter='t' where id = 53597; > > UPDATE 1 > > > > SELECT count(1) from transactions_sco where traiter='f'; > > count > > ------- > > 17589 > > > > AS YOU CAN SEE TWO ROWS WHERE UPDATED INSTEAD OF ONE !! > > THE COMMON THINGS BETWEEN THE TWO ROWS IS THAT THE FIELDS: cod_etu, > > cursus_id,vers_id,code_type_academic are the same > > Because that's what you asked upd1/2 to do for you. To see what is > happening, try selecting row id=53597 then manually running each rule > yourself, substituting in the OLD.foo from your selected row. You should > find that there are two rows that match 53597 on (cursus_id, vers_id, > traiter, code_type_academic) - itself and one other. Sorry, I didn't understand the manuel test procedure What is happening here? I am doing an update and the condition is on the ID and it is corresponding to the last Rule so why should the other rules interfer. Thanx for your help Najib. ---------------------------(end of broadcast)--------------------------- TIP 8: explain analyze is your friend
