[sqlalchemy] Re: Opinion on correct use of Sqlalchemy
On Apr 9, 2007, at 10:06 PM, Huy Do wrote: > > Michael Bayer wrote: >> Particularly for your query you are doing an eager load between >> "asset" and "location" yet a lot of your query criterion depends upon >> "location", so in that sense yes you have to use custom SQL, since >> query() will never involve eager loaded joins in the query criterion. > > Hi Michael, > > Everything you say makes perfect sense for 1:N relationships, but > in my > case, and with alot of other cases where I need the order by or the > criteria/filter on the joined table, it's a 1:1. In these cases I'm > not > sure why SA can't generate the same type of SQL statement that I am > above. It would make perfect sense for it to. > > I understand the eagerload problem with a list of child objects but > with > 1:1 relations I think the query interface should be querying in the > same > way that my manual SQL is. sorry, i just dont think that the loader strategy should *ever* have any effect on the primary results...otherwise loader strategies become intertwined with querying, and mappers become brittle since you can no longer change eager/lazy loads without breaking the results of your query. additionally for the eager loaders to guess when they should create non-aliased criterion and when they should create aliased criterion would be magical and complicated, and render improvements to eager loading impossible since people would be structutring their queries off of a now expected behavior which could then never change. it also implies that a lazy-loading mapper setup would break when eager loading is enabled via options, since the eager loader "expects" that its going to be used in query criterion, doesnt apply its aliases, and then changes the results of the query. so it definitely breaks a lot of widely accepted behavior. there are tools to join eager loads with query criterion which are contains_eager() and instances() or explcit select() objects passed to query.select(). id favor adding new mapper options to generate criterion that is simlar to that of the eager loader (such as, using contains_eager() with a Query-compiled query will add the columns in and the join for you) but loader strategies by default will never purposefully inject themselves into application query criterion. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Opinion on correct use of Sqlalchemy
Michael Bayer wrote:
> On Apr 7, 2007, at 8:11 AM, HD Mail wrote:
>
>
>> Hi,
>>
>> I was after some opinions on the following use of SA.
>>
>> 1. Is there any problems using SA in this way ?
>> 2. Is there better ways of achieving this ?
>>
>>
> ...
>
>> query = db.query(model.Asset).options(contains_eager('location'),
>> contains_eager('type'))
>> r = query.instances(s.execute())
>> return r, count
>>
>>
>
> that youve constructed exactly the query you want and used it via
> instances() is exactly how i want people to use SQLAlchemy when they
> know what non-trivial SQL they want to issue. query() only creates a
> very specific kind of SQL and it could never support figuring out how
> to construct the SQL that you already know how to do.
>
> Particularly for your query you are doing an eager load between
> "asset" and "location" yet a lot of your query criterion depends upon
> "location", so in that sense yes you have to use custom SQL, since
> query() will never involve eager loaded joins in the query criterion.
>
> however, theres a reason query follows this behavior, which is that
> if you are loading Asset objects with an eager loaded list of
> Location objects, but you have placed limiting criterion on the list
> of Locations specifically, you will now have Asset objects whose
> loaded list of Locations may not be complete compared to whats in the
> database (and they will remain that way until those instances are
> expire()d or similar). So you should be aware that that is the
> effect youre getting in your code.
>
Hi Michael,
Everything you say makes perfect sense for 1:N relationships, but in my
case, and with alot of other cases where I need the order by or the
criteria/filter on the joined table, it's a 1:1. In these cases I'm not
sure why SA can't generate the same type of SQL statement that I am
above. It would make perfect sense for it to.
I understand the eagerload problem with a list of child objects but with
1:1 relations I think the query interface should be querying in the same
way that my manual SQL is.
> also the separate "count()" step may be problematic since consider it
> wont return just the number of Asset objects loaded, but the number
> of rows total, which is Asset * Location * AssetType object rows. if
> you want just the number of Asset's returned youd just return len(r).
>
You're right, but because the the joins are 1:1, len(r) and count() will
give me the same result.
Thanks
Huy
> >
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: problems with threadlocal strategy and ResultProxy
id very much prefer a test case that isnt using Twisted since Ive no
idea what twisted does with threads. it looks like the result proxy
is being shared between threads (which i believe is an effect of the
twisted reactor anyway?).
if youre playing with threads to that degree the "threadlocal"
strategy may be too simplistic.
On Apr 9, 2007, at 5:41 PM, Manlio Perillo wrote:
>
> Hi.
>
> I have found some problems with threadlocal strategy and the
> ResultProxy.
>
> Here is the code that reproduces the problem, Twisted and
> PostgreSQL are
> required.
>
> I'm using SQLAlchemy 0.3.6-1 from Debian Etch.
>
> I don't know if this is a bug of SQLAlchemy.
> Is the ResultProxy thread safe?
>
>
>
> #! /usr/bin/env python
> import sys
>
> from twisted.python import log
> from twisted.internet import defer, threads, reactor
>
> import sqlalchemy as sq
>
>
> def callback(result, i):
> # Is the ResultProxy thread safe?
> if USE_CURSOR:
> rows = list(result)
>
> log.msg('callback', i)
>
> def errback(reason):
> log.err(reason)
>
> def f():
> l = []
>
> for i in range(N):
> log.msg('request', i)
> d = threads.deferToThread(db.execute, QUERY)
> d.addCallbacks(callback, errback, callbackArgs=(i,))
> l.append(d)
>
> return defer.DeferredList(l)
>
> def main():
> return f().addCallbacks(lambda _: None, lambda r: log.err(r)
> ).addBoth(lambda _: reactor.stop())
>
>
> POOL_SIZE = 1
> N = 10
> QUERY = 'SELECT CURRENT_TIMESTAMP'
> URL = 'postgres://manlio:[EMAIL PROTECTED]/test'
> STRATEGY = 'threadlocal'
> USE_CURSOR = False
>
>
> reactor.suggestThreadPoolSize(POOL_SIZE)
> db = sq.create_engine(URL, pool_size=POOL_SIZE, strategy=STRATEGY,
>echo_pool=False)
>
> log.startLogging(sys.stdout)
> reactor.callWhenRunning(main)
>
> reactor.run()
>
>
>
> When USE_CURSOR is False, I got (at random) the error:
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 718,
> in execute
> connection = self.contextual_connect
> (close_with_result=True)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
> line 148, in contextual_connect
> return self.session.get_connection(**kwargs)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
> line 21, in get_connection
> return TLConnection(self,
> close_with_result=close_with_result)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
> line 63, in __init__
> base.Connection.__init__(self, session.engine,
> close_with_result=close_with_result)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 368,
> in __init__
> self.__connection = connection or engine.raw_connection()
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
> line 127, in raw_connection
> return self.connection_provider.get_connection()
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/default.py", line
> 19, in get_connection
> return self._pool.connect()
>File "/usr/lib/python2.4/site-packages/sqlalchemy/pool.py",
> line 150, in connect
> return
> self._threadconns[thread.get_ident()].connfairy().checkout()
> exceptions.AttributeError: 'NoneType' object has no attribute
> 'checkout'
>
>
> When USE_CURSOR is True, I obtain (at random) any of these two errors:
>File "./concurrency.twisted.py", line 12, in callback
> rows = list(result)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 921,
> in __iter__
> row = self.fetchone()
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 997,
> in fetchone
> self.close()
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 862,
> in close
> self.cursor.close()
>File "/usr/lib/python2.4/site-packages/sqlalchemy/pool.py",
> line 312, in close
> if self in self.__parent._cursors:
> exceptions.TypeError: iterable argument required
>
>
> and:
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 719,
> in execute
> return connection.execute(statement, *multiparams,
> **params)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 444,
> in execute
> return Connection.executors[c](self, object,
> *multiparams,
> **params)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line
> 458,
> in execute_text
> cursor = self._execute_raw(statement, parameters)
>File
> "/usr/lib/python2.4/site-packages/sqlalchemy/engine/
[sqlalchemy] Re: handing SELECT ... FOR UPDATE
Jonathan LaCour wrote: > What would be the best way to approach using `SELECT ... FOR UPDATE`? Answering my own question, it appears that there is a `lockmode` kwarg you can pass when creating a query that supports 'update' and 'update_nowait' as values: query = session.query(Queue, lockmode='update') Just to make sure: is this what I need to be using, or am I missing something? I found this by digging through the source since its not mentioned in the docs, and it looks like it does what I want. -- Jonathan LaCour http://cleverdevil.org --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] problems with threadlocal strategy and ResultProxy
Hi.
I have found some problems with threadlocal strategy and the ResultProxy.
Here is the code that reproduces the problem, Twisted and PostgreSQL are
required.
I'm using SQLAlchemy 0.3.6-1 from Debian Etch.
I don't know if this is a bug of SQLAlchemy.
Is the ResultProxy thread safe?
#! /usr/bin/env python
import sys
from twisted.python import log
from twisted.internet import defer, threads, reactor
import sqlalchemy as sq
def callback(result, i):
# Is the ResultProxy thread safe?
if USE_CURSOR:
rows = list(result)
log.msg('callback', i)
def errback(reason):
log.err(reason)
def f():
l = []
for i in range(N):
log.msg('request', i)
d = threads.deferToThread(db.execute, QUERY)
d.addCallbacks(callback, errback, callbackArgs=(i,))
l.append(d)
return defer.DeferredList(l)
def main():
return f().addCallbacks(lambda _: None, lambda r: log.err(r)
).addBoth(lambda _: reactor.stop())
POOL_SIZE = 1
N = 10
QUERY = 'SELECT CURRENT_TIMESTAMP'
URL = 'postgres://manlio:[EMAIL PROTECTED]/test'
STRATEGY = 'threadlocal'
USE_CURSOR = False
reactor.suggestThreadPoolSize(POOL_SIZE)
db = sq.create_engine(URL, pool_size=POOL_SIZE, strategy=STRATEGY,
echo_pool=False)
log.startLogging(sys.stdout)
reactor.callWhenRunning(main)
reactor.run()
When USE_CURSOR is False, I got (at random) the error:
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 718,
in execute
connection = self.contextual_connect(close_with_result=True)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
line 148, in contextual_connect
return self.session.get_connection(**kwargs)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
line 21, in get_connection
return TLConnection(self, close_with_result=close_with_result)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
line 63, in __init__
base.Connection.__init__(self, session.engine,
close_with_result=close_with_result)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 368,
in __init__
self.__connection = connection or engine.raw_connection()
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/threadlocal.py",
line 127, in raw_connection
return self.connection_provider.get_connection()
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/default.py", line
19, in get_connection
return self._pool.connect()
File "/usr/lib/python2.4/site-packages/sqlalchemy/pool.py",
line 150, in connect
return
self._threadconns[thread.get_ident()].connfairy().checkout()
exceptions.AttributeError: 'NoneType' object has no attribute
'checkout'
When USE_CURSOR is True, I obtain (at random) any of these two errors:
File "./concurrency.twisted.py", line 12, in callback
rows = list(result)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 921,
in __iter__
row = self.fetchone()
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 997,
in fetchone
self.close()
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 862,
in close
self.cursor.close()
File "/usr/lib/python2.4/site-packages/sqlalchemy/pool.py",
line 312, in close
if self in self.__parent._cursors:
exceptions.TypeError: iterable argument required
and:
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 719,
in execute
return connection.execute(statement, *multiparams, **params)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 444,
in execute
return Connection.executors[c](self, object, *multiparams,
**params)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 458,
in execute_text
cursor = self._execute_raw(statement, parameters)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/engine/base.py", line 540,
in _execute_raw
cursor = self.__engine.dialect.create_cursor(self.connection)
File
"/usr/lib/python2.4/site-packages/sqlalchemy/databases/postgres.py",
line 277, in create_cursor
return connection.cursor()
File "/usr/lib/python2.4/site-packages/sqlalchemy/pool.py",
line 253, in cursor
self.invalidate()
File "/usr/lib/python2.4/site-packages/sqlalchemy/pool.py",
line 243, in invalidate
raise exceptions.InvalidRequestError("This connection is
closed")
sqlalchemy.exceptions.InvalidRequestError: This connection is
closed
Thanks and regards Manlio Perillo
--~--~-~--~~-
[sqlalchemy] handing SELECT ... FOR UPDATE
So, at work, we have a particular use case where we need to be able to do a `SELECT ... FOR UPDATE` in order to lock some rows. In this particular project we will be using the ORM package and some advanced data mapping (including multiple table polymorphic inheritance). What would be the best way to approach using `SELECT ... FOR UPDATE`? We were thinking about the possibility of creating a MapperExtension to make this happen, but aren't really sure about how to go about implementing that, and any direction would be useful. Thanks! -- Jonathan LaCour http://cleverdevil.org --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Selecting objects with null foreign keys
Suppose I have a table/object called Node that maps to Host by relation field 'host' on foreign key host_id that may be null. Why can't I do the following: Query(Node).select_by(host=None) When it seems obvious that I mean this: Query(Node).select_by(Node.c.host_id==None) I think the former would be cleaner, since I could just pass a 'host' variable that may be None without having to check it manually first before getting its id (however simple that might be). Thanks. Ryan --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Dealing with undefined/NULL foreign keys
Thanks thats brilliant.
Brendan
On 4/9/07, Michael Bayer <[EMAIL PROTECTED]> wrote:
>
>
> On Apr 9, 2007, at 2:38 PM, Brendan Arnold wrote:
>
> >
> > Hi there,
> >
> >> of course theres another view of this, which is why does your
> >> application prefer a "blank" Address entry to None? the latter is
> >> more semantically correct.
> >
> >
> >
> > Name:
> > ${person.name}
> > City:
> > % if person.address:
> > % city = person.address.city
> > % else:
> > % city = 'Not specified'
> > ${city}
> >
>
> gotya. the None version is like:
>
> ${person.address and person.address.city or "Not Specified"}
>
> >
>
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: Dealing with undefined/NULL foreign keys
On Apr 9, 2007, at 2:38 PM, Brendan Arnold wrote:
>
> Hi there,
>
>> of course theres another view of this, which is why does your
>> application prefer a "blank" Address entry to None? the latter is
>> more semantically correct.
>
>
>
> Name:
> ${person.name}
> City:
> % if person.address:
> % city = person.address.city
> % else:
> % city = 'Not specified'
> ${city}
>
gotya. the None version is like:
${person.address and person.address.city or "Not Specified"}
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: Dealing with undefined/NULL foreign keys
Hi there,
> of course theres another view of this, which is why does your
> application prefer a "blank" Address entry to None? the latter is
> more semantically correct.
I'm exporting the Person object to a web-template and I prefer the following,
Name:
${person.name}
City:
${person.address.city or 'Not specified'}
to the following,
Name:
${person.name}
City:
% try:
% person.address.city
% except NameError:
% person.address = Address(city='Not specified')
${person.address.city}
or, perhaps better
Name:
${person.name}
City:
% if person.address:
% city = person.address.city
% else:
% city = 'Not specified'
${city}
Cheers!
Brendan
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: Dealing with undefined/NULL foreign keys
On Apr 9, 2007, at 11:25 AM, Brendan Arnold wrote:
> Is there a way I can ensure that an empty instance of my Address
> object is grafted onto the Person object even when corresponding data
> is not found in the people table?
you can either create a layer of abstraction between the actual
mapped property and the publically accessible field by using a property:
class User(object):
def _get_address(self):
if self._address is None:
self._address = Address()
return self._address
address = property(_get_address)
mapper(User, users, properties={"_address":relation(Address)})
or you can use a MapperExtension and override create_instance():
class MyExt(MapperExtension):
def create_instance(self, mapper, selectcontext, row, class_):
instance = User()
if not row['address_id']:
instance.address = Address()
return instance
mapper(User, users, properties=..., extension=MyExt())
in both cases, flushing the User object will create a new address row
in the DB since you have the blank address object there...to avoid
that, you might have to add more layers of abstraction to the public
"address" property.
of course theres another view of this, which is why does your
application prefer a "blank" Address entry to None? the latter is
more semantically correct.
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: Putting Session.flush in a seperate thread
On Apr 9, 2007, at 11:54 AM, David Anderson wrote: > > On 4/9/07, Michael Bayer <[EMAIL PROTECTED]> wrote: >>> I tried to build this already but I cannot figure out how to assign >>> the shared connection to a thread. I always get 'No connection >>> defined'. >> >> if youre on py2.4 check out threading.local(). > > Just to be clear here, as I'm also having issues with SQA and > threading : a connection is per-thread, so if I have 10 worker threads > that each read/write to the db, I need to manually create 10 > connections and pass them into each thread? What would the elegant way > to do this be? the elegant way would be to not worry about the thread scope of a connection and just use connection pooling normally. pull out a connection when needed, close it when complete (which is really just a return to the connection pool). keep the scope of those two operations local to a single thread. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Support for ordered lists of child items
we dont have the capability to automatically update ordering columns when the elements of a list are moved around. if you move the elements around, you need to execute some step that will update the index columns (or create a custom collection class that does this for you). On Apr 9, 2007, at 12:42 PM, Aaron Digulla wrote: > > Hello, > > I'm looking for a feature but couldn't find it in the docs. > > I have a tree like structure where the user can specify the order of > the children of a node. In DB lingo, I have a parentId and an index > column. When I load children, they should be ordered by the index. > This seems to be supported. > > Can SA also update the index column when I move children in the list > around? Like: > > # ... parent has three children A, B C > item = parent.children[0] > del parent.children[0] > parent.children.insert (1, item) > # now, parent has three children B, A, C > > Regards, > > > > --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Support for ordered lists of child items
Hello, I'm looking for a feature but couldn't find it in the docs. I have a tree like structure where the user can specify the order of the children of a node. In DB lingo, I have a parentId and an index column. When I load children, they should be ordered by the index. This seems to be supported. Can SA also update the index column when I move children in the list around? Like: # ... parent has three children A, B C item = parent.children[0] del parent.children[0] parent.children.insert (1, item) # now, parent has three children B, A, C Regards, --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Putting Session.flush in a seperate thread
On 4/9/07, Michael Bayer <[EMAIL PROTECTED]> wrote: > > I tried to build this already but I cannot figure out how to assign > > the shared connection to a thread. I always get 'No connection > > defined'. > > if youre on py2.4 check out threading.local(). Just to be clear here, as I'm also having issues with SQA and threading : a connection is per-thread, so if I have 10 worker threads that each read/write to the db, I need to manually create 10 connections and pass them into each thread? What would the elegant way to do this be? Thanks, - Dave --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Dealing with undefined/NULL foreign keys
Hi there,
I have a table of, say, people and another table of addresses. the
people table refers to addresses through a foreign key. It is optional
for each entry in the people table to have an address.
When I construct an object using
assign_mapper(session_context, Person, people_table, properties = {
'address' : relation(Address) })
and the foreign key in the people table is NULL (i.e. the address is
not specified) a None object is grafted onto the Person object instead
of an Address object. This causes all-sorts of errors later in the
code.
I remedied this by including an entry in the addresses table which
corresponds to an unspecified address, but this feels wrong somehow
(i.e. when listing addresses, need to parse out the 'unspecified'
entry etc.)
Is there a way I can ensure that an empty instance of my Address
object is grafted onto the Person object even when corresponding data
is not found in the people table?
Regards,
brendan Arnold
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: Putting Session.flush in a seperate thread
On Apr 9, 2007, at 5:18 AM, Koen Bok wrote: > > We are building a GUI app, and we were thinking about wrapping > session.flush() in a thread with a timer that detects a timeout. That > way we would have better performace and we can generate warnings if > the connection goes down. Do you guys think this is smart, or are > there complications? > we have logic for invalidating connections that have disconnected in the trunk, but it relies upon the DBAPI throwing exceptions. session.flush() is a very high level place to put connection timeout checking, also what about selects/queries ? > I tried to build this already but I cannot figure out how to assign > the shared connection to a thread. I always get 'No connection > defined'. if youre on py2.4 check out threading.local(). --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Putting Session.flush in a seperate thread
may be a threadlocal strategy. On 4/9/07, Koen Bok <[EMAIL PROTECTED]> wrote: > > > We are building a GUI app, and we were thinking about wrapping > session.flush() in a thread with a timer that detects a timeout. That > way we would have better performace and we can generate warnings if > the connection goes down. Do you guys think this is smart, or are > there complications? > > I tried to build this already but I cannot figure out how to assign > the shared connection to a thread. I always get 'No connection > defined'. > > > > > --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Ordering by field in related object
Thanks a lot!! Grin the Django query object actually allows for this sort of ordering so i just figured SqlAlchemy should allow it as well :/ But never mind now I know the trick it's super easy to adjust for it ;) --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Putting Session.flush in a seperate thread
We are building a GUI app, and we were thinking about wrapping session.flush() in a thread with a timer that detects a timeout. That way we would have better performace and we can generate warnings if the connection goes down. Do you guys think this is smart, or are there complications? I tried to build this already but I cannot figure out how to assign the shared connection to a thread. I always get 'No connection defined'. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
