Marco Gaiarin a écrit:
we have decided recently to go ... to a deepest mySQL dependency. mySQL has a feature that doesn't exist in ANSI-SQL92 : full-text indexing.
http://www.postgresql.org/idocs/index.php?functions-matching.html#FUNCTIONS-...
No, but pg, as often in GNU programs, have regexp. ;)
OK, but can pg use indexes on select with regexp ? Because the problem is not on "can it be done", but "can it be done FAST", to avoid a 5-10 seconds waiting time on each biblio select !
I think it is not the case to fight against one or other DBMS, simply wrote down a correct relational database schema in normal form, and put inside as standard as possible SQL instructions, for the rest using specialized instructions case()ing on the DBMS type.
I agree with you for the schema in a normal form. Let's release a stable 2.0 and improve this aspect (with many others...) -- Paul POULAIN Consultant indépendant en logiciels libres responsable francophone de koha (SIGB libre http://www.koha-fr.org)