Hey gang -
A feature I have procrastinated on for weeks is now complete -
deferred property loading ! and as a related bonus, minimal column
updates !
Deferred Property Loading
------------------------------------
deferred property loading is basically "lazy load for scalar
properties". heres how it looks:
m = mapper(Order, orders, properties={
'description':deferred(orders.c.description)
})
Now, when you load some Order objects in, the "description" column
will not be in the select list. it will only be loaded in when you
actually access it. This is a really key memory/performance feature
- for examples like the "blog demo", the front page listing of posts
would normally load the full row in for each entry into memory,
including the "body" which could be very large, even though its
never displayed. By "deferring" it you save the trouble of loading
it all the time and only when its actually needed.
Minimal Column Updates
-----------------------------------
The saving of objects that results in an UPDATE statement has also
been modified to only specify SET clauses for those columns that
have actually changed. This keeps "deferred" columns from being
loaded unnecessarily, and also keeps potentially large numbers of
fields or large-sized fields from being repeatedly updated with the
same data.
Deferral Groups
-----------------------
A 'group' string can be specified to group multiple properties
together into a single deferred load, triggered when you first access
any of the grouped properties:
m = mapper(PressRelease, pressreleases, properties = {
'body' : deferred(pressreleases.c.body, group='secondary'),
'image' : deferred(pressreleases.c.image_data,
group='secondary'),
'credits':deferred(pressrelease.c.credit, group='secondary'),
})
release = m.get_by(release_id=7)
# when you access 'body', 'image', or 'credits', all three will be
loaded in
print "body: " , release.body
Deferred Options
------------------------
The "deferred" idea can also be swtiched on or off or modified with
the options 'defer' and 'undefer':
# get a copy of 'm' that loads 'description' in the main query
m2 = m.options(undefer('description'))
# get a copy of 'm' that defers the loading of 'data' in
m3 = m.options(defer('data'), undefer('description'))
# add a property to a 'deferral group'
m4 = m.options(defer('data', group='load_info'))
# property names can also be . separated to specify nested mappers
dmapper = myclass.mapper.options(defer('prop1.prop2.largefield'))
Hope everyone's testing is going well.
- mike
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Sqlalchemy-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/sqlalchemy-users
