something with lazy-relations, self-referential, but pointing to
another instance, not to _itself_. Happens regardless of inheritance.
2308 is all-ok, 2344 is same as 2309.
hmm, what should i do to this all-combination test, so it gets
included it in the testsuite? i.e. make it unittest-like
And another one - at same revision - maybe same thing?
IMO the UPDATE for a lazy post_update relations gets the dbapi
values swapped, e.g. (sqlite)
rev2308:
SA: INFO UPDATE Person SET friend_id=? WHERE Person.db_id = ?
* SA: INFO [2, 7]
rev2309-2344:
* SA: INFO UPDATE Person SET
sorry, on more file needed
something with lazy-relations, self-referential, but pointing to
another instance, not to _itself_. Happens regardless of
inheritance.
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
I personally use latest python
On 2/17/07, Ali [EMAIL PROTECTED] wrote:
Yes, I imagined something like that. I only posted to avert users from
norcebo effect that they might encounter reading this thread.
On Feb 17, 5:16 pm, Michael Bayer [EMAIL PROTECTED] wrote:
if you didnt have to
I'm getting the following error when I build the select clause.
select_columns = [table.c.col1, table.c.col2]
statement = table.select(select_columns)
File build\bdist.win32\egg\sqlalchemy\sql.py, line 1351, in select
File build\bdist.win32\egg\sqlalchemy\sql.py, line 65, in select
I'm new to SQLAlchemy, and am currently using it for Turbogears
developement. I've run into an issue calling delete on an instance. This
almost exact same code works fine in one of my projects, but I get an error
in this particular one. I think it may have be because this is a
many-to-many
select_columns = [table.c.col1, table.c.col2]
statement = table.select(select_columns)
You probably want to:
select_columns = [table.c.col1, table.c.col2]
statement = sqlalchemy.select(select_columns)
res = statement.execute().fetchall() # or the like...
The table.select form you try to
Ok. I need to do the following:
select col1, col2, sum(col3), sum(col4) where col1 = test group by
col1, col2
Q1: How do I add sum functions in the column list
Q2: Do I add group_by using the group_by parameter on the
sqlalchemy.select call
Q3: Do I add where conditions to the select using
I cant run the generator since theres a dependency Config not
found, but the standalones are using the foreignkey parameter which
is both deprecated and also more or less mutually exclusive with
remote_side.
the trunk now has the foreign_keys parameter which behaves
consistently as a
this is something related to how turbogears is integrating
SQLAlchemy. there might be a bug in SA here but I cant track it down
without an example that shows what TG is doing more specifically. in
particular i hope they are using the QueuePool connection pool and
not the
yeah actually foreignkey's legacy behavior is fine, those examples
are using it exactly incorrectlythe parameter takes on a
different meaning when used in a self-ref relationship (thats why its
been replaced) - its synonymous with remote_side in a self-ref
relationship.
On Feb 20,
seems the following will do the job:
group_columns = [table.c.col1, table.c.col2]
select_columns = group_columns + [sqlalchemy.func.sum(table.c.col3),
sqlalchemy.func.sum(table.c.col4)]
statement = sqlalchemy.select( select_columns, (table.c.col1 ==
test), group_by=group_columns)
res =
foreignkey is out. dont use it at all.
remote_side, when the remote side of the relationship is ambiguous
based on the join condition, i.e. a self referential mapping
relationship. can be provided for otherwise non-ambiguous cases as
well.
foreign_keys, when your join condition contains
Hi,
I am using the recipe for multiple foreign keys to the same table at
http://www.sqlalchemy.org/trac/wiki/UsageRecipes/MultiFkRelations
I am also using SessionContext, and assign_mapper
It works well, except for one issue. The second object (ie that which
the second foreign key references)
On Feb 20, 12:01 pm, Ali [EMAIL PROTECTED] wrote:
Here is my code (with other fields removed):
stockreceipt_table = Table('stockreceipt', meta,
Column('user_id', Integer, ForeignKey('user.id')),
# Cancellation
Column('cancelled_user_id', Integer, ForeignKey('user.id')),
)
I left out the fields for brevity of example. They both have an id
column defined as primary keys.
On 20 Feb, 17:11, Michael Bayer [EMAIL PROTECTED] wrote:
On Feb 20, 12:01 pm, Ali [EMAIL PROTECTED] wrote:
Here is my code (with other fields removed):
stockreceipt_table =
By both, I mean user and receipt tables
On 20 Feb, 17:12, Ali [EMAIL PROTECTED] wrote:
I left out the fields for brevity of example. They both have an id
column defined as primary keys.
On 20 Feb, 17:11, Michael Bayer [EMAIL PROTECTED] wrote:
On Feb 20, 12:01 pm, Ali [EMAIL PROTECTED]
using the foreignkey parameter
which is both deprecated and also more or less mutually exclusive
with remote_side.
specifying foreign_keys explicitly, _and_ remote_side does work.
here the last fixed variant.
but if u comment out foreign_keys and lave only remote_side, in cases
like B
Matt Culbreth [EMAIL PROTECTED] writes:
I'm playing out with a few things now and I wanted to see if anyone
else has used SQLAlchemy in an asynchronous manner? For example, you
could create a service which responded to asynchronous requests for
data, and could be used by a web client,
the current error message is:
raise exceptions.ArgumentError(Cant determine
relation direction for relationship '%s' - foreign key columns are
present in both the parent and the child's mapped tables. Specify
'foreign_keys' argument. %(str(self)))
are u on trunk ?
Just a quick question, but is there code available to autogenerate code for
SA or Elixir from an existing (rather large) database? I could use SQLSoup
directly, but it really helps me to be able to see the generated code.
Also, some of the classes will benefit from some added methods.
Thanks,
On Feb 20, 12:18 pm, David Bolen [EMAIL PROTECTED] wrote:
Michael Bayer [EMAIL PROTECTED] writes:
I suppose in this case where I've got a better handle on the object
actually being changed (the one being saved()), that expunging that
one object might be the most effective approach, and is
Ok,
from sqlalchemy import *
from sqlalchemy.ext.assignmapper import assign_mapper
from sqlalchemy.ext.sessioncontext import SessionContext
meta = DynamicMetaData()
ctx = SessionContext(create_session)
def make_engine(uri=None):
if uri is None:
uri = 'sqlite:///:memory:'
well I cant run the program since its not complete (no Drug, Supplier
class/tables defined, etc), but if id hazard a guess id say you
shouldnt be calling flush() on your class instances, since it will
not flush any changes on any dependencies. call ctx.current.flush()
instead.
On Feb
Thanks for the advice, and sorry I didn't provide the correct
information, but ctx.current.flush() makes no difference.
r.cancelled_user is (nearly: 19/20) always None immediately after the
assign and flush with no database access being made. Interestingly,
the user_id/user relationship is always
Instead of isouter argument to Join, have join_type. Default to
'inner' but could also be 'left', 'right', or 'outer.' (Or even
'natural' but that might be confusing b/c of Join's default behavior
of joining by FK which is similar to but not the same as natural
join.)
(Of course, it's quite
On 2/20/07, Jonathan Ellis [EMAIL PROTECTED] wrote:
Instead of isouter argument to Join, have join_type. Default to
'inner' but could also be 'left', 'right', or 'outer.'
should read ... or 'full,' of course.
--~--~-~--~~~---~--~~
You received this message
ive no problem with that feature, it was a TODO for sometime that got
whacked from the sql.py module. feel free to patch this up, though
its a lot of work - to get it working on Oracle in ANSI mode using
(+), maybe get DB's that dont support FULL OUTER JOIN to raise an
error, etc.
Hi,
I have performance issues when trying to upload a large amount of data
to my Oracle db using SQLAlchemy's ORM.
My DBA told me to add a hint /*+APPEND*/ to my insert DML statment,
but I couldn't find a way in SQLAlchemy to add a hint to the DML
generated by the ORM.
I searched google but to
29 matches
Mail list logo