> > Is there any reason that it can't be split up into logical units eg > > gmVaccinations.sql, gmAllergies.sql etc?. > > No reason. However, if so do maintain the separation of > static (tables) vs dynamic (pretty much everything else) > stuff. The reason is that the dynamic stuff can be *rerun* > on an existing database without putting the actual data in > jeopardy at all.
is there a reason comments can't be kept with tables. I would have thought that makes someone trying to learn by reading the sql code much easier for them. > > > Some of the areas causing me problems seem to be around statements > > containing lines like: > > > > alter table clin.vacc_def (from gmClinicalViews.sql) > > add constraint numbered_shot_xor_booster > > check ( > > ((is_booster is true) and (seq_no is null)) or > > ((is_booster is false) and (seq_no > 0)) > > ); > > > > this won't compile, stops and asks to input a parameter > > Very strange. I tend to think pgAdmin is not sophisticated > enough. The above is perfectly valid SQL and it bootstraps > just fine. You might have to contact the pgAdmin people. This is interesting Karsten, I'll do some more research on it. The problem lies after the check part of the query. I've found that some statements containing 'check' compile without problems, yet others don't - perhaps there is something in the text after it. BTW I'm not using pgAdmin3, but datamanager from www.sqlmanager.net/ , you can download a free version, but I paid them as I find it useful. The professional version for winxp is reasonably stable, and the visual query builder is fabulous. The linux version is unfortunately not very good, so I use pgadmin when I'm in linux. Regards richard _______________________________________________ Gnumed-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnumed-devel
