occasionally, i run into this problem with Postgres; though it would likely
happen on other systems too.
my database might miss a migration, and we end up with this situation
SqlAlchemy Model - Column = varchar(1000)
PostgreSQL - Field = varchar(255)
Data = 500 characters in length
When I try
i think it is from the DBAPI. i was (sadly) expecting you to say that.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sqlalchemy+unsubscr...@googlegroups.com.
To
Michael-
Quick question for clarity...
I have a table with a few deferred columns. If I want to eagerly load
them during a query, I should pass in the undefer option right ?
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from
that's not too bad. at least for me; i only defer a few HSTORE columns.
On Thursday, May 15, 2014 2:24:51 PM UTC-4, Michael Bayer wrote:
yeah there’s some bad history with the API here, in that you can’t easily
pass them all at once, but right now it’s q.options(undefer(‘*’)) for
I'm not sure if there is a trick to do this or not, but it's worth asking...
I have an object that has been loaded into the Session, as per whatever
eager-loading requirements :
foo = dbSession.query( Foo ).filter(...).options(...).one()
I'm now in a position where I need to access
thanks for the writeup. i'll tinker with some ideas in my downtime. for
now, i'll stick with the secondary , non-Foo, query I have that pulls that
bar/baz stuff.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and
the easiest trick i came up with for this, was to create a logging table
in sqlalchemy/postgres and directly log the query into it.
my table looks roughly like this:
id , timestamp, request_id ( if under pyramid ), query_id (if available,
guid assigned to different queries ), query, vars,
if i have any time after shipping , i'll probably build in transaction
support for celery and pyramid.
I keep away from tossing ORM objects around the system. GETS are pretty
cheap.
my task arguments are generally:
int = primary key of ORM object
dict = instructions payload of what
i'll try this later.
i ended up 'monkeypatching' all these relationships onto the classes at the
end of my models.py file , using the non-string column syntax. would have
preferred to keep the entire class definition, together.. but I documented
everything.
--
You received this message
I just ran into an issue where it looks like I could have race conditions
using SqlAlchemy and Celery. Wondering if anyone here has some ideas.
Here's the scenario:
A1 Process A - Pyramid - Creates SQLalchemy session.
A2 Process A - Pyramid - Creates data
A3 Process A - Pyramid - flushes data
i've got that now as a stopgap; i was hoping someone has better ideas. i
don't like the idea of a post-commit hook, because i fear requesting the
celery task request will create an error. I really don't want to build
`transaction` support for celery, but i might need to.
--
You received
I'm on postgres and have a boolean column that allows NULL values.
I need to create a relationship between 2 ORM classes , where there is a
filter that states IS NOT TRUE.
The ORM likes these 2 commands :
photo = sa.orm.relationship(Photo, primaryjoin=and_(
Useraccount.photo_id==Photo.id
I had a similar situation years ago. We had software that helped automate
online promotions for music releases. Everyone insisted on keeping their
data separate; forcing different databases was required by contract and we
were strong-armed into it. Today, things would be different.
I ended
I have these 2 tables in the database:
class TableA(base):
id
field_a
field_b
items = sa.orm.relationship(TableA_Items, primaryjoin=TableA.id==
TableA_Items.table_a_id,
backref=table_a)
class TableA_Items(base):
id
table_a_id
thank you !!
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sqlalchemy+unsubscr...@googlegroups.com.
To post to this group, send email to
sorry for overload the list with random questions. i'm trying to get a
release out the door and dealing with performance bottlenecks...
given this setup:
class Items2Attributes(base):
id
item_id
condition_id
attribute_id
class Items(base):
id
On Friday, April 4, 2014 8:11:11 PM UTC-4, Michael Bayer wrote:
yeah, make one relationship and just filter them with a @property for each
of “a” and “b”. if the object is usually used such that both collections
are needed, that’s the approach.
brilliant. thanks!
--
You received
I'm calling this eagerloading , but it's not the right term...
I have an object that is already loaded...
user = dbSession.query( model.User ).filter_by( id = 1 ).first()
over time it's seen a handful of relationships grow...
user = dbSession.query( model.User ).filter_by( id = 1 )\
I'd like to avoid that, if at all possible. that was actually a solution to
an earlier thing i had a few weeks ago!
I'm fine with doing suqueryloads instead of joinedloads or hitting the
database multiple times.
with my current design... building another query where i need to specify
loads
If you have:
class MyTable(base):
pass
You could do
session.query( MyTable ).from_statement()
*
http://docs.sqlalchemy.org/en/rel_0_9/orm/query.html#sqlalchemy.orm.query.Query.from_statement
See also:
http://docs.sqlalchemy.org/en/rel_0_9/core/tutorial.html#using-text
ah!
i didn't catch this before...
backref_name = timeseries_%s % name ## might be better as an the id
of the location
location = relationship('Location',
backref=backref(backref_name, lazy='dynamic'))
I don't think he can do one table. originally i did, but I looked into some
I'm interested in what you find. I know TypeDecorator is the right
solution, I was looking at that for this exact same situation a few weeks
ago
(
https://groups.google.com/forum/#!searchin/sqlalchemy/vanasco%7Csort:date/sqlalchemy/sQtOYxSUiqI/5ns2vWMFaGAJ
)
I have a similar situation. I
I have a handful of queries that are fairly complex.
For these, I need to support both
SELECT object.* ( for full records )
SELECT object.id ( for caching of collection ids + COUNT operations )
My current approach is really dirty. three functions wrap a core
generator, which
No. I'm want to edit the target of the 'select' from an existing query.
For many select operations, I need three variations
- the full Objects
- just the ObjectIds
- a count of the objects ( which is sometimes faster as an explicit select
obj.id than a select obj , so i can influence the pg
If you can only use FTP, then your options is probably just to write a
program that downloads the sqlite via ftp, does the queries, then uploads.
that's going to be messy with race conditions though, so you'd have to
create some sort of system where the file is renamed/hidden, downloaded,
collection = one to many / many to many.
scalar relationship = one-to-one or many-to-one relationship.
when defining a relationship, you pass in 'uselist=False' to create define
the relationship as scalar
--
You received this message because you are subscribed to the Google Groups
The data comes in unordered and sometimes contains duplicates, so there's
a UniqueConstraint on Entry on sub, division, created.
Have you tried pre-processing the list first ?
I've had similar situations, when dealing with browser , user and app
analytics.
I normally do a first pass to
That's a really simple use case -- you want the primary key to be a 1:1
relationship, and you don't have potentially competing fields.
But this gets complicated when:
- You want User.addresses to be a single item ( User.address )
- You don't want User.addresses to just be all the addresses --
Since you're using Postgres... have you considered using python to generate
a COPY file ?
Sqlalchemy doesn't seem to support it natively... maybe via 'text', but
your underlying psycopg2 driver does.
it's way way way faster. i've found it significantly faster than dropping
fkeys and using
i'm updating my db code, and this popped up:
InvalidRequestError: One or more mappers failed to initialize - can't
proceed with initialization of other mappers. Original exception was: id()
takes exactly one argument
is there any way to tell where/how this got triggered? There's nothing
thanks. i finally found it.
i was querying TableA
TableA mapped somewhere to TableB
Table B had a backref on an 'id' field. that field got removed.
the query on TableA triggered the error , even though I wasn't loading the
'id' field. because of how the backref was constructed, it was
You would probably do better with a pattern where you have a Session for
every request , and just use that session. That is how most people
implement SqlAlchemy for web.
* request start
* create a sqlalchemy session, either scoped or explicit
* do things with your session : read ,
Before writing stored procedures, you might want to test speeds like :
1. raw speed within the mysql client for these queries.
2. speed of pushing these queries through your python dbapi driver ( ie,
mysqldb )
those should give you a better idea on where the potential speed problems,
if any
ok. i guess i'll experiment with listeners to see if I can grab and store
the original object into a dict on load, and then compare against that.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving
I'm using twisted and looking to move raw SQL ( via twisted's
`runInteraction` ) into my SqlAlchemy model. I've only played with it a
bit.
A few notes:
Under Twisted you can't have threads spawned threads. Your computer will
want to die. I did and crashed the server several times,
I have some lightweight revision tracking on some models. I generate a
diff based on the history of the object (via inspector).
a limitation I just realized, is that this history only dates back to the
most recent flush() -- it doesn't date back to the initial load.
are there any existing
On Monday, March 3, 2014 5:26:29 AM UTC-5, Simon King wrote:
Partially innocent question: If one of the flushes you describe above
fails
(say with an integrity error), then the transactions would be left in
need
of a rollback on all engines and definitely no data would be
If this behavior is a bug , I can build out a testcase. FWIW , I think
it's a bug.
class Foo :
id = int primary key
bar = relationship(bar)
baz = relationship(baz)
# hits the db
a = session.query( Foo ).get(1)
# doesn't hit the db
b =
On Friday, February 28, 2014 12:41:46 PM UTC-5, Michael Bayer wrote:
nope. get() means, “the SELECT may not happen”. eager loads are always
secondary to the fact that the row is being loaded. if you want to
definitely load the row, say filter_by(id…).first().
Poking at the source
On Friday, February 28, 2014 2:33:11 PM UTC-5, Michael Bayer wrote:
OK well it has to emit the SELECT for the primary row, but yes
populate_existing will overwrite, so really, just any equivalent of
query(User).filter_by(primary key).first() is really all you need here…we
used to have a
On Tuesday, February 25, 2014 5:07:11 AM UTC-5, Simon King wrote:
However, Python has a handy shortcut for passing arbitrary keyword
arguments to a function:
status_args = {trans_status: 1}
status = Status(**status_args)
You could also do that in one line, but i'm sure it violates
If you do need to use zope.transaction ( and not just the SqlAlchemy option
that Simon posted ) , there already exists Zope/SqlAlchemy integration --
https://pypi.python.org/pypi/zope.sqlalchemy
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To
To expand and clarify Simon's answer into bulletpoints :
Expired Object
* The database Session the object belongs to has been closed ( or committed
unless you tweaked the config )
* SqlAlchemy considers it 'expired' because it is not reasonable to expect
the object to reflect the current state
quick question on the definitive behavior for the orm
I'm have some code that reassigns objects from one owner to another.
instead of objects being assigned directly to an owner, they're assigned
into an intermediary Library, which is then assigned to an Owner
i also keep the owner_Id on the
Wonderful. Thanks a ton.
Seriously. Thanks for answering all these super-simple questions. I'm
doing some code-audits and last-minute prep for release. There have been a
handful of simple things that just seemed to work ( this, the dogpile
cache, etc ) , and I want to confirm that
I'd suggest recompiling mysqldb
I've had weird/odd issues happen when either Python or MySQL/PostgreSQL
were updated and the driver wasn't. Even for minor updates ( Python 2.7.4
- 2.7.5 ; Postgres 9.3 - 9.4, etc ), the compiled c-extension will often
cause random errors if anything that
i currently have a table that looks like this:
class MyTable(DeclaredTable):
id = sa.Column(sa.Integer, primary_key=True)
features_enabled = sa.Column(sa.Integer, nullable=False, default=0,
)
_feature_manager = None
@property
def
that's fine. this is for a webapp where we have in a single request :
begin;
user = .filter().first()
DO LOTS OF STUFF, all over the place
DO EVEN MORE STUFF , in more places
user = .get(user_id)
commit;
if this behavior is intended, then we can just rely on it for now.
Wow! Thanks. This is by far, the best written and more informative stuff
I've read on this.
I've been considering switching to AWS sooner-than-later for the RDS ( we
bootstrapped on Linode with a plan to redeploy on AWS in the future ) , but
it looks like we shouldn't pursue that yet!
On
I've had problems with this in the past. I thought it was SqlAlchemy, and
it ended up being issues with my data in Pyramid (the Pylons successor). A
few items had the wrong character encoding -- I was expecting it to be
UTF8/ASCII, but it was using a different charset.
It could be the
I don't know if any of this will help - I'm too tired to read though all
your code:
1. Twisted isn't threadsafe, so anything happening in deferToThread needs
to be entirely encapsulated. This has totally screwed me up many times.
You should avoid a scoped_session , as it has to do with
Couldn't it be handled with a mixin?
{{{
class _LoadCore(Base):
whatever you want for both classes here
pass
class Load(_LoadCore):
__tablename__ = 'load'
__mapper_args__ = {
'polymorphic_identity':'load',
'polymorphic_on':'polymorphic_type',
}
id =
just wondering -- would it be possible to mimic this behavior using a
custom view for this select ( in postgresql ) and then querying that ? i
think the query planner might only run on the creation.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy
On Thursday, February 6, 2014 2:18:51 PM UTC-5, Klauss wrote:
I think views don't cache the plan, they're handled as rules.
What you'd need is a function (pg's version of stored procedures).
I had time to look it up; this generally seems correct. Looking at some
explain syntax, it seems
you can pass custom connect arguments to the DBAPI in `create_engine`.
http://docs.sqlalchemy.org/en/rel_0_9/faq.html?#how-do-i-pass-custom-connect-arguments-to-my-database-api
depending on which driver you use, it will be different.
you shouldn't use `text`, because sqlalchemy
just to clarify...
mysqldb - `connect` accepts a `sql_mode` string:
http://mysql-python.sourceforge.net/MySQLdb.html
pymysql accepts it too
https://github.com/PyMySQL/PyMySQL/blob/3576863f9cd0b66ce6c8b32ab3448ab68f55f489/pymysql/connections.py
oursql:
i couldn't find it
When you called 'commit', the 'account' object got expired.
There are a few ways in the FAQ or Wiki to deal with this. IIRC, you can
`merge` objects back into the session or `refresh` the object. One or both
might hit the database.
It would probably be better to change your code a bit, so
Check out this project:
https://pypi.python.org/pypi/sqlacodegen
it's a replacement to sqlautocode
the packages are designed to inspect your database and create sqlalchemy
models ( in python code) for you.
you'll have to edit/audit the generated code -- but it will save A LOT of
time.
--
i believe you need to be explicit on your group_by :
group_by(P.id)
or
group_by(P.attrs)
or
group_by(P.id,P.attrs)
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it,
Just to rephrase everyone else's It doesn't matter response:
Many ORMs out there REQUIRE the database be built by the ORM or designed by
the ORM. This is because the ORM stores and accesses data in a very
particular manner -- so tables and columns must adhere to certain naming
conventions,
Would an association proxy help ?
http://docs.sqlalchemy.org/en/rel_0_9/orm/extensions/associationproxy.html
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
i'm just wondering for performance over-the-wire from the bandwidth alone...
I see great, DBA toubleshooting friendly stuff like:
SELECT table_one.column_one AS table_one_column_one FROM table_one ;
but that could be...
SELECT t1.column_1 AS t1_column_1 FROM table_one t1;
or even
so wow. that's really neat. amazingly neat.
and yeah, i'm just looking for ways to get some more performance out of a
few boxes , so we don't have to add another box. trying to cut the fat here
and there -- and noticed some very verbose sql. wondering if losing it will
get 1-2% more out of
sorry , should have been more clear. i'm trying to get some more juice out
of of the database server. it is streaming sql nonstop. the webservers
are doing fine, and are simple to cluster out.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
i couldn't find anything on trac earlier for `coerce_from_config`
i'd be happy to whip up a patch for the .8 branch that
updates coerce_from_config to support all the create_engine kwargs. i have
no idea how to do that for the .9 branch though.
my sql is pretty optimized as is, with
if the pessimistic listener doesn't work, maybe he could do something like
this:
* create an event listener to grab the sql executes
* wrap the execute in try/except with a savepoint
savepoint()
while True :
try:
execute()
break
except:
rollback()
fails
i needed to do some savepoint work in sqlalchemy, and remembered mike
directed me to a FAQ or example with a contextmanager syntax:
with session.begin_nested():
pass
i jumped on the sqlalchemy docs, and couldn't find that at all.
the orm docs (
automap sounds neat! thanks!
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sqlalchemy+unsubscr...@googlegroups.com.
To post to this group, send email to
You can also lock down the database to read-only --
* create Read-Only database users/privileges if you have an authenticated
database
* if you're on something like sqlite, you can run your process as different
users -- so one can access the DB for read and other for write.
--
You received
On Saturday, December 14, 2013 9:58:14 PM UTC-5, Jeff Dairiki wrote:
Oh my god. I'm just going to close that page, and try to forget I
ever saw that... On quick look though, none of those modes seem to
have anything to do with transactions (thankfully).
Yeah, it's pretty evil. I've
just a thought--
two things:
1- The thing is, for me, if the session has only been used for read
operation,
self.tx seems to be None. So the datamanager never commits anything.
Check out Tres Seaver's reply
here: https://groups.google.com/d/msg/pylons-discuss/R4S-UwHV6ww/ekD7M9UEvp8J
richard - how are you handle the scoping and management of sessions ?
i'm wanting (badly) to convert some twisted code that is using raw sql
through their db api onto the sqlalchemy model for my core app.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy
On Friday, December 6, 2013 10:16:21 AM UTC-5, Sunil Adapa wrote:
Before making a connection attempt the overflow counter lock is obtained
and it is being released only after the connection either succeeds or
fails. In my case, a connection remained hung possibly because of a surge
in
can you share your existing schema for these relations ?
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sqlalchemy+unsubscr...@googlegroups.com.
To post to this
what about using custom compiler functions
(
http://docs.sqlalchemy.org/en/rel_0_9/core/compiler.html#utc-timestamp-function
) ?
you could make a custom function...
sum_integer()
and just call that instead of the default sqlalchemy option
--
You received this message because you are
I just learned all about the Events model. You could potentially write a
listener that can filter the data for you. I'm not very familiar with the
events though, and it could be a bit difficult to do -- as some items would
be functions, others column operations, and others from 'text'
this is a really great idea! thanks for asking this question.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sqlalchemy+unsubscr...@googlegroups.com.
To post to
i'm looking at moving some raw sql in twisted to SqlAlchemy and have a
question.
I have a multi-threaded twisted daemon that tends to generate a lot of race
conditions on a few tables that are frequently hit.
I get integrity errors from something like this :
domain = SELECT * FROM
oh that's great - I didn't expect SqlAlchemy to aggregate/support the
different driver errors like that!
thanks so much, Michael!
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it,
On Monday, December 2, 2013 2:46:20 PM UTC-5, Michael Bayer wrote:
well the alternative to full outer join is a union of two outer joins -
more tedious.“full outer join” isn’t in the library but you can just
subclass Join and do a @compiles to get that exact syntax.
you could also
I don't like this schema.
1. It makes more sense to have an 'contact_id' as a column on Company
Account , than vice versa. A contact is a representative for those
entities, not the other way around. I think you're running into issues,
because you seem to be trying to adapt this real-world
On Sunday, November 24, 2013 5:56:29 PM UTC-5, Joseph Casale wrote:
On Sunday, November 24, 2013 3:44:32 PM UTC-7, Michael Bayer wrote
conditonals?
Yeah, the queries are just so long and with all the combinations of
possible
criteria it would get out of hand as I have about 6 optional
for what it's worth...
usually I see errors in psycopg and other c-extended python libraries when
the python and/or postgres versions got upgraded. it can get tricky when
you're using a virtualenv, because you need to rebuild the entire
virtualenv - and sometimes people forget that. ( i know
On Monday, November 18, 2013 8:40:13 PM UTC-5, Michael Bayer wrote:
there’s a recipe for doing this with events:
http://www.sqlalchemy.org/trac/wiki/UsageRecipes/SessionModifiedSQL
Oh wow. This is great. Playing with it now.
I still like my idea for syntax though , it's simpler and
Thats a great idea. I'll tackle that next.
I'm having a bit of trouble adapting the recipe to my existing [pyramid]
application.
I'm not seeing a .info property on my sessions in
`_connection_for_session`.
I am using scoped sessions, and the recipe does work fine on this machine /
how stable is 0.9 ? is it fairly safe to switch to now ?
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sqlalchemy+unsubscr...@googlegroups.com.
To post to this
screw it. updated to 0.9. everything works as-needed within 5 minutes.
happy.
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
is the HSTORE column inheriting from MutableDIct ?
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sqlalchemy+unsubscr...@googlegroups.com.
To post to this group,
forgot to add -- this creates the following SQL:
UPDATE user_data SET prefs=(user_data.prefs || hstore(ARRAY[%(param_1)s,
%(param_2)s])) WHERE user_data.prefs ? %(prefs_1)s
{'prefs_1': 'inboundBusStop', 'param_1': 'transitAvaiable', 'param_2':
'true'}
Note that it's using one of the htsore
i remember getting annoyed with this a while back. I fought with
SqlAlchemy and Postgresql to get the right SQL generated ( Postgres had a
few bad lines in the docs that confused things ).
I ended up NOT being able to get this to work using the ORM. However I
found a workaround using the
On Monday, November 18, 2013 1:46:18 PM UTC-5, Michael Bayer wrote:
s.query(A).update({A.data: A.data + {foo: bat, hoho: lala}},
synchronize_session=fetch)
...
the UPDATE looks like:
UPDATE a SET data=(a.data || %(data_1)s)
{'data_1': {'foo': 'bat', 'hoho': 'lala’}}
that's a whole
this is just wild thinking -- and might not be possible.
I'm wondering if anyone has some ideas on creating a relationship where
only certain columns are queried during relationship loading.
I have a few relationships like this ( pseudocode )
class Tag2Posting:
tag_id INT
Something that I realized would be very useful, is to enable comments on
query compilation. That would make it SO much easier to understand logs.
Yes this sounds slightly silly.
Yes I am 100% serious.
I've been working all day on optimizing a web page. It originally had 200
queries,
Wouldn't selecting the sql logger and logging the comment provide you
the same information?
You'd only be able to push a comment into Python's log, but not into
Postgres/Mysql/Oracle/etc.
placing a comment into the sql allows one to grep the database logs, or use
the database profiling.
Well the queries are totally different...
SELECT * FROM test WHERE id IN (SELECT test.id FROM test LIMIT 10)
That's 100% in your database, and a single executable.
SELECT * FROM test WHERE id IN ( 1...10 )
You have SQL Alchemy generating that query all the bind params, then
You can do something like:
q1 = session.query(\
ContactInfo.account_id.label(q1_account_id ),
ContactInfo.other_id.label(q1_otherid )
)\
.filter(..)
q2 = session.quiery(Site.account_id.label(q2_account_id)).filter(..)
_q3 =
sidenote: there are typos. that's just to illustrate. i mostly copied
that from production code.
the core concepts are:
1. `label()` the columns
2. use the `engine` (not orm) to join the queries
3. `alias` the union to search against them
4. the labeled columns in the earlier queries are now
In my experience, you need to write a script that you can configure
as-needed. Depending on the types of inserts, I've had different
performances.
* I always set up batches to be configurable ; I start at 100, then try
to go up down
* I set toggles/thresholds on when I `flush`. After every
i think the call to alias ( making it a subquery ) dropped the support for
all in my example; though that was written in .8.1 -- might be working
since .8.3 or 9
it took me a bit of trial and error to get stuff working.
awesome that it's working for you so quickly!
--
You received this
You missed adding the object to the session, before the flush.
this should work:
new_foo = Foo('red')
dbSession.add(new_foo)
dbSession.flush()
print new_foo.id
--
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To unsubscribe from
801 - 900 of 1035 matches
Mail list logo
