I'd recommend keeping the writes to a single thread if you can as that will
be easiest. If you can't do that then you could tie them together using the
transaction module's two phase commit and zope.sqlalchemy. As you want the
same transaction shared across threads you'll want to create your
I've been experimenting with precompiled queries in my app using the second
recipe
on https://bitbucket.org/zzzeek/sqlalchemy/wiki/UsageRecipes/BakedQuery.
I'm seeing a big speed up, with pages 30-40% faster (my application has
graph structured content so some pages require ~100 items.) But
IIRC, the implementation of the two phase commit basically has everyone
flush on Phase1, then report back as a vote. If there are any negative
votes, the transaction manager instructs everyone to fail and rollback as
Phase2. If you had a flush that caused an integrity error before
On Friday, 10 January 2014 08:06:16 UTC-8, Michael Bayer wrote:
maybe. I was thinking, is the Session really doing the right thing here
by not getting involved, but I think yeah that still might be appropriate.
so we’d need to carry along some extra information about the transaction
On Thursday, 9 January 2014 09:41:40 UTC-8, Jeff Dairiki wrote:
Okay, I've traced things out a bit more.
If the session state is not STATUS_INVALIDATED (aka STATUS_CHANGED),
SessionDataManager.commit() does a self._finish('no work'). That is
where self.tx gets set to None (this ---
On Monday, 23 December 2013 06:38:41 UTC-8, Michael Bayer wrote:
On Dec 23, 2013, at 9:29 AM, Sibylle Koczian nulla.e...@web.dejavascript:
wrote:
Am 21.12.2013 16:27, schrieb Michael Bayer:
In the case of using Postgresql, the type
sqlalchemy.dialects.postgresql.INTERVAL takes
On Thursday, 12 December 2013 15:26:08 UTC-8, Thierry Florac wrote:
Hi,
I'm using two-phase transactions with ZODB, PostgreSQL and Oracle
databases connected with SQLAlchemy without problem.
I'm not using native zope.sqlalchemy package, but another package called
ztfy.alchemy that I've
On Thursday, 12 December 2013 16:30:59 UTC-8, Jeff Dairiki wrote:
Do you understand why the datamanager is finding the SessionTransaction
and using that directly? (At least I think that's what it's doing ---
I haven't sussed this out completely.) I'm referring to the line from
On Friday, 18 February 2011 08:14:28 UTC-8, Michael Bayer wrote:
we've put tickets on their tracker to this effect, that they should be
more liberal about considering when the transaction begins.
http://code.google.com/p/pysqlite/issues/detail?id=21
pysqlite is tricky since I dont know if
I'd like to measure the number of database round trips that are associated
with a request to my web application so I can write tests to catch
potential performance regressions. I've been using
mock.Mock(wraps=connection.execute) to keep count and while I think this
works for selects, I don't
On Tuesday, 26 March 2013 14:54:08 UTC-7, Michael Bayer wrote:
On Mar 26, 2013, at 5:24 PM, Laurence Rowe lauren...@gmail.comjavascript:
wrote:
I'd like to measure the number of database round trips that are
associated with a request to my web application so I can write tests
I've just uploaded the 0.5 release of zope.sqlalchemy to PYPI,
bringing SQLAlchemy 0.6 support.
http://pypi.python.org/pypi/zope.sqlalchemy
Savepoint release support (nested transaction commit) is currently
being discussed on ZODB-dev:
On Apr 28, 4:38 pm, Chris Withers ch...@simplistix.co.uk wrote:
Laurence Rowe wrote:
Chris,
This is what the combination of repoze.tm2/transaction and
zope.sqlalchemy does for you. You don't have to do anything special
other than that.
It doesn't do the .remove().
BFG currently has
Chris,
This is what the combination of repoze.tm2/transaction and
zope.sqlalchemy does for you. You don't have to do anything special
other than that.
Laurence
On Apr 28, 2:37 pm, Chris Withers ch...@simplistix.co.uk wrote:
Hi All,
I'm still trying to get an answer on this...
Am I right in
Hi,
Following a request[1] for savepoint release support in
zope.sqlalchemy, I've been looking into how this might be done. Adding
the necessary support to Zope's transaction module was quite simple
[2], but the mapping of Zope transaction savepoints - SQLAlchemy
nested transactions - database
On Jan 16, 6:39 pm, Michael Bayer mike...@zzzcomputing.com wrote:
On Jan 16, 2010, at 12:35 PM, Laurence Rowe wrote:
Hi,
Following a request[1] for savepoint release support in
zope.sqlalchemy, I've been looking into how this might be done. Adding
the necessary support to Zope's
It would be interesting to see if this could be made to work. The
SimpleDB model is rather different from the relational model, so it
would only be useful if your application does not use any advanced
features - no joins etc, each 'domain' might map to one big (albeit
sparse) table.
Laurence
On
On Nov 5, 9:14 am, Martijn Faassen [EMAIL PROTECTED] wrote:
Michael Bayer wrote:
I wonder if a before_ hook should be provided as well.
I don't know myself; Laurence, if you're listening in perhaps you'd care
to comment?
While there is aesthetic value to having symmetrical before and
On Sep 23, 4:20 pm, mg [EMAIL PROTECTED] wrote:
Hello There,
I am developing an application that uses sqlalchemy and the py
processing packages. My question is this, what is the best practice
for using sessions in this type of app. Each subprocess needs to
access my db to get work, so
19 matches
Mail list logo