[orm-devel] new many2one relationship in 0.9pre
Tue, 18 Feb 2003 01:53:12 +0100
I've added a new relationship/column pair called
brought up my Mike, thought through and named by Ross and implemented
by myself this is a good example of teamwork ;-))
I moved some stuff from the one2one relationship into a common
baseclass for 2one relationships which made more sense in terms of
design than Mike's idea of just subclassing one2one, which would have
worked, but this looks nicer.
If there are no major complains I'll raise orm some attention on
freshmeat tomorrow with the 0.9 release.
available now and I updated CVS repository, too.
The changes have brought up an old problem I can see with orm:
caching. If you look at my test_many2one.py and run it you'll see that
orm performes one SELECT on the author table for each book although
most of them are redundant. I have intentionally not implemented a
caching mechanism of some sort, because I still want people to buy the
next version ;-) What I'm really saying is, caching, views and the oid
issue should wait untill after 1.0.
JOINs are another thing I'd like to take a look at. With the tables
from my example you'd
SELECT book.id,book.title,author.id,author.name FROM book,author
Although this retrieves a lot of redundant data from the db it is
actually better than orm's current solution: fewer (just one) SELECT
request. I wonder how orm should do this. I have to think about this
Any input is much appreciated!
Good night for me...
_..._ Diedrich Vorberg
/ _ _ \ http://www.tux4web.de
| (o)_(o) | firstname.lastname@example.org
\( ) / .---.
//'._.'\ \ / \ Internet Dienstleistungen
// . \ \ \.@-@./ und 'Consulting'.
|| . \ \ /`\_/`\
|\ : / | // _ \\ Linux Rules!
\ `) ' (` /_ | \ )|_
_)``".____,.'"` (_ /`\_`> <_/ \
) )'--'( ( \__/'---'\__/