Hi Mario, You are right. The processing of your JPQL query does not go through the method I suggested. Instead, it is handled in classes from the org.apache.openjpa.jdbc.kernel.exps package - mainly Count class and UnaryOp.appendTo method. You could probably modify the appendTo method to suit your need. However, I guess you don't want to modify the OpenJPA source code but provide some extension only. This might be impossible for these classes, at least I don't know the way to do it.
My other idea then is to look at getSelects and toSelect methods in DBDictionary. Hope this helps, Milosz > > > I had tried it already, Milosz. > But this overwritten method doesn't seem to be running. I debugged it, > others overwritten methods runs. > > > > Miłosz Tylenda wrote: > > > > Mauro, > > > > I would start with looking at DBDictionary.toSelectCount method. > > > > Regards, > > Milosz > > > > > >> The jpql: "SELECT count(o) FROM Object o" (A) > >> > >> is usually translated to this sql: "SELECT count(t.primarykey) FROM Table > >> t" > >> (B) > >> > >> I'm working with a legacy database witch jdbc driver doesn't accept this > >> format of sql. > >> It works with: "SELECT count(*) FROM Tabelat" (C) > >> > >> Is there a way to make the query C (with count(*)) be generated instead > >> of > >> query B? > >> > >> I'm using the default jpa dictionary. I know that I can extend > >> DBDictionary > >> in order to customize this. > >> I had some difficulty trying to do this. > >> > >> Somebody would give some hint? > >> > >> Thanks. Mauro. > >> > >> -- > >> View this message in context: > >> http://n2.nabble.com/COUNT-t-to-COUNT-tp4176827p4176827.html > >> Sent from the OpenJPA Users mailing list archive at Nabble.com. > > > > > > > > -- > View this message in context: > http://n2.nabble.com/COUNT-t-to-COUNT-tp4176827p4182603.html > Sent from the OpenJPA Users mailing list archive at Nabble.com. >
