On Nov 10, 4:57 am, Petr Kobalíček [EMAIL PROTECTED] wrote:
I have postgres related problem. I'm normally developing with sqlite,
but we are using postgres on the server. The problem is that
sqlalchemy probably remembers primary keys and after database restore
it will start in all tables
Hi!
I use PostgreSQL and when I try to create schema I use following
command conn.execute('CREATE SCHEMA %(s)s', {'s':s})
I get raise exc.DBAPIError.instance(statement, parameters, e,
connection_invalidated=is_disconnect)
sqlalchemy.exc.ProgrammingError: (ProgrammingError) syntax error at or
Hi,
you could simply get the Bowler objects from the results by saying
results = [r[0] for r in results]
I'm not sure whether you query is correct, though. Usually, you cannot
select columns which are not in the GROUP BY clause or which are not
aggregated - after grouping, you have several
Hi,
I am using SQLAlchemy with PostgreSQL 8.3.
Today I just discovered the great postgres_returning functionnality.
Dumb question :
Is it possible to take advantage of it in session mode ? In other
words, is it possible for session.add() to issue only one INSERT
INTO ... RETURNING ... query
Thanks Simon - just checked and I'm running 2.5.2 on my machines.
From experimenting - I'm not so sure I have a memory leak, so much as
just using a lot of memory.
I didn't realise that when Python frees memory, it doesnt necessarily
become free in Linux. I think that possibly all that's
Hello Alchemy Land!
If I have a simple test-case with Bowler objects and City objects, and
I want to use func.max and group_by in order to find the highest
scorers in each city... I might do something like this:
max_score = func.max(Bowler.highscore).label('highest_score')
results =
Note that a lot of database drivers cache *everything* in memory when
you .fetchall(), fetchone() or fetchmany(x). So all those operations
consume the same amout of memory :
result = cursor.execute(...)
for i in result:
...
data = result.fetchall()
for i in data:
...
data =
On Mon, Nov 10, 2008 at 4:33 AM, King Simon-NFHD78
[EMAIL PROTECTED] wrote:
I'm no SQL expert, so please take this with a pinch of salt, but as far
as I know, conditions in the 'WHERE' clause of an SQL statement are
applied BEFORE any grouping, so you can't use grouping functions (such
as
-Original Message-
From: sqlalchemy@googlegroups.com
[mailto:[EMAIL PROTECTED] On Behalf Of John Hunter
Sent: 10 November 2008 14:07
To: sqlalchemy@googlegroups.com
Subject: [sqlalchemy] Re: select where field=max(field)
On Mon, Nov 10, 2008 at 4:33 AM, King Simon-NFHD78
Hi,
Apologies for lowering the general IQ of the list, I'm very new to web apps
and databases.
I had a declarative table:
class ArkContact(Base):
table of all contacts
__tablename__ = 'contacts'
id = Column(Integer, primary_key=True)
project_id = Column(Integer,
hi:
In my app there is a user_table, with a column access_time.
Without sqlalchemy, just update user_table set access_time =
Now() ,
With sqlalchemy and user as a object-relational object, I have to
make a app time and do user.access_time = now() ?
or a better way?
On Nov 10, 2008, at 5:35 AM, Ian Charnas wrote:
Hello Alchemy Land!
If I have a simple test-case with Bowler objects and City objects, and
I want to use func.max and group_by in order to find the highest
scorers in each city... I might do something like this:
max_score =
Hi all,
I'm using SA 0.5.0rc3 and MySQL 5.0.51a on Mac OS X 10.4.11. I have a
table with a float column and would like to have a default value of 0:
Column('col', Float(), default=0.0)
However, executing metadata.create_all(engine) yields
CREATE TABLE `Table` (
...
`col` float default
The new behavior is exactly what I expect, namely that query.count()
returns the same as len(query.all()). Are there cases in which this
does not make sense or where this would not work?
-- Christoph
--~--~-~--~~~---~--~~
You received this message because you are
-Original Message-
From: sqlalchemy@googlegroups.com
[mailto:[EMAIL PROTECTED] On Behalf Of joelanman
Sent: 10 November 2008 00:21
To: sqlalchemy
Subject: [sqlalchemy] Re: Memory leak - is session.close() sufficient?
Thanks for all the advice - I've changed my unicode settings
On Nov 10, 2008, at 2:25 AM, Randall Smith wrote:
Just to make sure we're considering the same plan, I don't plan to
make
any API changes that would cause breakage. All changes are additions
including the public API and some new dialect methods (get_views,
get_indexes, ...). Most of
Simon wrote:
Hi all,
I'm using SA 0.5.0rc3 and MySQL 5.0.51a on Mac OS X 10.4.11. I have a
table with a float column and would like to have a default value of 0:
Column('col', Float(), default=0.0)
However, executing metadata.create_all(engine) yields
CREATE TABLE `Table` (
...
Postgres doesn't allow bind parameters to be used with CREATE SCHEMA -
it expects an identifier, not a literal value.When I try it on my
system I don't get the E behavior you're getting, I get :
sqlalchemy.exc.ProgrammingError: (ProgrammingError) syntax error at or
near 'foo'
LINE 1:
On Nov 10, 2008, at 12:08 PM, John Hunter wrote:
On Mon, Nov 10, 2008 at 10:05 AM, King Simon-NFHD78
[EMAIL PROTECTED] wrote:
Actually, the section after that (Using Subqueries) probably does
something very close to what you want. What's the result of these
lines:
q1 =
Michael,
Michael Bayer wrote:
I know what this is and it should be working in r5280. I don't have
access to firebird here so we weren't able to run the tests on it
before rc3 was out.
Thanks for the quick reply.
Looking at the changes doc these will be included in rc4 - any idea
I know what this is and it should be working in r5280. I don't have
access to firebird here so we weren't able to run the tests on it
before rc3 was out.
On Nov 10, 2008, at 7:39 AM, Werner F. Bruhin wrote:
I am getting sometimes the following exception with rc3 which I did
not
see
On Nov 10, 2008, at 12:10 PM, Werner F. Bruhin wrote:
Michael,
Michael Bayer wrote:
I know what this is and it should be working in r5280. I don't have
access to firebird here so we weren't able to run the tests on it
before rc3 was out.
Thanks for the quick reply.
Looking at the
On Mon, Nov 10, 2008 at 10:05 AM, King Simon-NFHD78
[EMAIL PROTECTED] wrote:
Actually, the section after that (Using Subqueries) probably does
something very close to what you want. What's the result of these lines:
q1 = (session.query(Snapshot.strategy, Snapshot.symbol, sum_pnl)
it should be fine.
On Nov 10, 2008, at 6:34 AM, Cito wrote:
The new behavior is exactly what I expect, namely that query.count()
returns the same as len(query.all()). Are there cases in which this
does not make sense or where this would not work?
-- Christoph
On Nov 10, 2008, at 8:37 AM, paftek wrote:
Hi,
I am using SQLAlchemy with PostgreSQL 8.3.
Today I just discovered the great postgres_returning functionnality.
Dumb question :
Is it possible to take advantage of it in session mode ? In other
words, is it possible for session.add() to
I am getting sometimes the following exception with rc3 which I did not
see with rc2 when I do something like this:
engine = sa.create_engine(dburl, encoding='utf8', echo=False)
# connect to the database
##connection = engine.connect()
Session = sao.sessionmaker()
-Original Message-
From: sqlalchemy@googlegroups.com
[mailto:[EMAIL PROTECTED] On Behalf Of John Hunter
Sent: 08 November 2008 05:09
To: sqlalchemy@googlegroups.com
Subject: [sqlalchemy] Re: select where field=max(field)
[SNIP]
Here is a query that lists the sum(pnl) for each
Simon, Michael, thank you!
Simon, yes you were totally right, my query was totally wrong! I was
up all night trying to get some code working, and at 5am I was getting
a little fuzzy. I'd like to use that as my excuse ;-)
What I ended up doing this morning was doing a simple query with max
and
With 0.4 it's a positional argument to Column: Column('col', Float(),
PassiveDefault('0.0'))
Simon wrote:
Thanks Jason! Is there any way of doing this in SA 0.4 as well?
On 10 Nov., 16:42, jason kirtland [EMAIL PROTECTED] wrote:
Simon wrote:
Hi all,
I'm using SA 0.5.0rc3 and MySQL
Had a problem this morning where SA 0.5rc3 was returning None, while
0.5rc2 and 0.4.8 returned the expected object/row.
tables mappers:
--
typehierarchy_table = Table('typehierarchy', metadata, autoload=True)
typehierarchy_names_table =
I sent this last week but it seems like it may not have been
posted to the list...at least, I haven't seen any responses :)
-- Forwarded message --
Date: Tue, 4 Nov 2008 18:00:57 -0500 (EST)
From: [EMAIL PROTECTED]
To: sqlalchemy@googlegroups.com
Subject: extending pre-existing
On Nov 10, 2008, at 3:42 PM, [EMAIL PROTECTED] wrote:
Right now I'm playing with mapper inheritance. The first stumbling
block I've come to is the case where the local table doesn't yet have
a row for the object from the base table. In that case, a query on my
subclassed object returns
Hi David -
One thing I notice is that your remote_side on the self referential
relation from TypeNode-Children is not needed, whereas it *is* needed
on the TypeNode-Parent side, which is the many to one side, using
backref=backref('Parent', remote_side=[typehierarchy_table.c.id]).
On Mon, 10 Nov 2008 at 17:19, Michael Bayer wrote:
Possibly. The fragility here is that you are relying on a model
that isn't actually implemented here, i.e. that your application is
written around a table inheritance assumption when that is not
actually the case - the extended tables may or
Yeah, I kinda figured that out in a round-about way, because 0.5rc2 and
0.4.8 would get me my object, but then node.Parent was the collection of
children. Thanks for the response.
Michael Bayer wrote:
Hi David -
One thing I notice is that your remote_side on the self referential
relation
I ended up having problems with my backref (in any version of SA) so I
re-worked my mappers and now 0.5rc3 is generating correct SQL. So this
probably isn't a bug.
mapper(TypeNode, typehierarchy_table, properties={
'AutoPopNames':relation(TypeAutoPop, backref='TypeNode'),
Traceback (most recent call last):
File /srv/server/metaserver/metaserver/lib/base.py, line 56, in
__call__
ret = WSGIController.__call__(self, environ, start_response)
File /usr/lib/python2.5/site-packages/Pylons-0.9.6.2-py2.5.egg/pylons/
controllers/core.py, line 195, in __call__
after =
On Nov 10, 2008, at 5:34 PM, [EMAIL PROTECTED] wrote:
On Mon, 10 Nov 2008 at 17:19, Michael Bayer wrote:
Possibly. The fragility here is that you are relying on a model
that isn't actually implemented here, i.e. that your application is
written around a table inheritance assumption when
On Mon, Nov 10, 2008 at 11:10 AM, Michael Bayer
[EMAIL PROTECTED] wrote:
you need an extra tuple on the join, query.join((q1, s.s==q1.c.s))
This gets past the syntax error, but does not produce the right
results. I had to take some time off today to work on other problems,
but am now
yeah this is an enhancement we made, whereby InstanceState removes
circular references from itself when its host object is garbage
collected, thereby taking the load off of gc (and it does). So in
this case, asynchronous gc is occurring right as InstanceState is
doing expire_attributes
gotcha, cool. was I first to run into this? :-)
On Nov 10, 5:57 pm, Michael Bayer [EMAIL PROTECTED] wrote:
yeah this is an enhancement we made, whereby InstanceState removes
circular references from itself when its host object is garbage
collected, thereby taking the load off of gc (and
Hi Chris,
yeah these tools works great, our problem was that if I did backup and
restoration from web interface then this problem happen.
I wasn't also familiar with postgres :)
Cheers
- Petr
2008/11/10 Chris Miles [EMAIL PROTECTED]:
On Nov 10, 4:57 am, Petr Kobalíček [EMAIL PROTECTED]
42 matches
Mail list logo