[Zope] Invalidate RAM cahce manager programatically
jim at ebizq.net
Wed Oct 15 09:53:34 EDT 2003
Ok. What I am doing is rather than get data from the DB directly, I
"cache" it in a Zcatalog. This give me the added side benefit of using
the ZCTextIndex's search capability for all content on my site as well.
How I do this is this: certain SQL methods use classes to massage the
result set into a loose class (there are no or few methods on these
classes). Then that class is passed to the Zcatalog with it's row_id as
the object identifier.
Here's the code sample that adds "feature" objects to the catalog (the
metadata_ids are unique and constant).
Python script catalogItem(metadata_id=0) exists in my feature ZCatalog
if metadata_id is not 0: <-- whoops. Just saw this funny, the code seems
to work though
features = context.sql.feature.getMetadata(metadata_id=metadata_id)
features = context.sql.feature.getMetadata()
for feature in features:
print "Cataloging metadata_id: %s<br>" % feature.metadata_id
# get my authors and their companies
for author in
print "Adding author: %s<br>" % author.name
uid = '%s' % feature.metadata_id
print "Added to catalog with ID: %s<br>" % uid
I do this for almost all other "content types" we have here with little
or no change. Could that goof with the "is not 0" be the cause for the
From: Chris Withers [mailto:chrisw at nipltd.com]
Sent: Wednesday, October 15, 2003 8:47 AM
To: Jim Kutter
Cc: zope at zope.org
Subject: Re: [Zope] Invalidate RAM cahce manager programatically
Jim Kutter wrote:
> What I discovered was that calling catalog_object doesn't re-catalog
> SQL results like I thought it would. I had to uncatalog_object first,
> then catalog_object to have the changes from the database kick in. I
> presumed that multiple calls to catalog_object would overwrite the
> ZCatalog's entry with new data.
They do, you must be doing something pretty weird for this not to work.
What does your code which calls catalog_object look like?
More information about the Zope