u can look at my posts on the topic and all related too (-:)
search for recursive/recursion/graph/sets
in short:
- storage:
- store graph in pure form (nodes/edges only) - easy to
maintain/update, very difficult to query
- store some preprocessed (cached) denormalized form (all paths
greetings,
i'm batch processing xml documents to mysql using SA 0.5, ORM.
data extracted from xml docs may be new or an update in form of a
replacement of
the existing object (record). (one of the columns, product_id, is
unique=True). Hence,
SA throws, as expected, an IntegrityError (1062,
After some further experimenting with this, it seems I cannot take a
mapped object I retrieved from one session and save it in another.
Using expunge then adding the object to a second session on a different
database does not work. I thought maybe it would.
This is a really useful thing to do,
read various threads about copy/deep-copy.
in any case u'll have to do the hierarchy-copying yourself - even if
moving one object (alone) from session to session succeeds somehow.
On Saturday 17 January 2009 14:47:41 Darren Govoni wrote:
After some further experimenting with this, it seems I
Yes, i'd like to do that, but because of the code injection seems to
carry state from the old session, the new session has trouble with even
the copy.
On Sat, 2009-01-17 at 15:23 +0200, a...@svilendobrev.com wrote:
read various threads about copy/deep-copy.
in any case u'll have to do the
I tried using session.merge(deepcopy) on my copy.deepcopy of a mapped
object(s) and when I commit the session I get
RuntimeError: maximum recursion depth exceeded
strange.
On Sat, 2009-01-17 at 15:23 +0200, a...@svilendobrev.com wrote:
read various threads about copy/deep-copy.
in any case
Yes, i'd like to do that, but because of the code injection seems
to carry state from the old session, the new session has trouble
with even the copy.
then, separate the issues:
a) make a working deep copy within same session (this is not at all
trivial, except for very simple schemas)
b)
Hi,
Did you first load the entry to update into the session or do you
always create new objects to be saved ?
I guess you need to:
1 - try to get the record from the database using the product_id or
another unique field
2 - if the result set is not empty,
update the record using
On Jan 17, 2009, at 4:07 AM, n00b wrote:
greetings,
i'm batch processing xml documents to mysql using SA 0.5, ORM.
data extracted from xml docs may be new or an update in form of a
replacement of
the existing object (record). (one of the columns, product_id, is
unique=True). Hence,
SA
thanks. i'll give it a shot.
On Jan 17, 10:43 am, Michael Bayer mike...@zzzcomputing.com wrote:
On Jan 17, 2009, at 4:07 AM, n00b wrote:
greetings,
i'm batch processing xml documents to mysql using SA 0.5, ORM.
data extracted from xml docs may be new or an update in form of a
actually, this is exactly what i didn't want to do .. at least not
explicitly. thx.
On Jan 17, 3:24 am, Laurent Rahuel laurent.rah...@gmail.com wrote:
Hi,
Did you first load the entry to update into the session or do you
always create new objects to be saved ?
I guess you need to:
1 -
On Jan 17, 2009, at 1:46 PM, Darren Govoni wrote:
Hi,
Thanks for the suggestions. In my system, it is a federation of
databases using SQLA as the object layer. There are message queues
that move detched data around as needed. My goal is to keep the
application as OO as possible with
This release includes a bunch of architectural enhancements that
didn't make it into 0.5.0, a handful of bug fixes, and one bigger bug
fix for a potential memory leak in Query. Upgrading is not mandatory
but is recommended.
Changes of note include:
- the memory leak bug is
I'm using SA underneath a TurboGears 1.0 app. Upgrading SA from 0.4.3
to 0.4.4 causes previously passing unit tests to fail when run in
conjunction with nose's coverage plugin -- I've included an example
stack trace below.
The unit tests run just fine when not using nose's --with-coverage
this is a known Python issue fixed in 2.6. See
http://www.sqlalchemy.org/trac/ticket/1138
for details.
On Jan 17, 2009, at 9:23 PM, James wrote:
I'm using SA underneath a TurboGears 1.0 app. Upgrading SA from 0.4.3
to 0.4.4 causes previously passing unit tests to fail when run in
15 matches
Mail list logo