Hi, all!
I use just one session in my project. Objects (loaded from the
database) can changed by user while the program in progress. But only
some of them should be saved. Usually I use session.flush
(selected_objects), but now this opportunity is deprecated :(
Please help me to find a solution
i want to add a composite index to the class inherited from
declarative_base
I tried this,
class MyClass:
__tablename__ = 'my_table'
id = Column(Integer, primary_key=True)
name = Column(String, nullable=False)
type = Column(String, nullable=False)
__table_args__ = (
On Aug 19, 2:27 am, King Simon-NFHD78 simon.k...@motorola.com
wrote:
-Original Message-
From: sqlalchemy@googlegroups.com
[mailto:sqlalch...@googlegroups.com] On Behalf Of darkblueB
Sent: 19 August 2009 02:58
To: sqlalchemy
Subject: [sqlalchemy] Re: new questions
Hi Simon
Has anyone generated ORM queries using the OVERLAPS SQL operator that
reference columns in the tables in the query? I've been experimenting
with various approaches and can't seem to cleanly get the column names
(with their appropriate alias based on the rest of the query) into the
overlaps
Faheem Mitha wrote:
Hi,
The following script is then followed by its output, and
finally by the
table output.
I don't get what is going on here. Yes, I should commit the
session, and
the table is empty as expected, but why does the id keep
incrementing on
successive runs, and
hi thanks for the reply..
but one doubt
how will i access the MyClass inside MyClass
i tried
ndex('ix_name_type',
MyClass.__table__.c.name,MyClass.__table__.c.type, unique=True)
it is giving the error
NameError: name 'MyClass' is not defined
thanks
On Aug 21, 2:04 pm, King Simon-NFHD78
i tried
class MyClass:
__tablename__ = 'my_table'
id = Column(Integer, primary_key=True)
name = Column(String, nullable=False)
type = Column(String, nullable=False)
__table_args__ = (
Index('ix_name_type', name , type ,unique=True)
)
it errors out
-Original Message-
From: sqlalchemy@googlegroups.com
[mailto:sqlalch...@googlegroups.com] On Behalf Of rajasekhar911
Sent: 21 August 2009 10:25
To: sqlalchemy
Subject: [sqlalchemy] Re: index in SA
i tried
class MyClass:
__tablename__ = 'my_table'
id =
exactly
On Aug 21, 2:33 pm, King Simon-NFHD78 simon.k...@motorola.com
wrote:
-Original Message-
From: sqlalchemy@googlegroups.com
[mailto:sqlalch...@googlegroups.com] On Behalf Of rajasekhar911
Sent: 21 August 2009 10:25
To: sqlalchemy
Subject: [sqlalchemy] Re: index in SA
This is something that could be improved in SQLAlchemy, but as a
workaround you can use the compiler extension to create the support
yourself. Here's some example code. It uses some private internals
from SQLAlchemy so you need to keep an eye on it that it doesn't break
when changing versions.
On Fri, 21 Aug 2009, King Simon-NFHD78 wrote:
I've never used postgres, but I believe auto-incrementing counters are
implemented using database sequences. I think these are incremented
outside of a transaction - this ensures that two seperate database
connections using the sequence at
Faheem Mitha wrote:
Thanks for the fast and helpful response. This looks like an artifact
of
how I am creating the table. I wonder if this would still show up if I
explicitly specified the id. I could check this. Also, presumably if I
had
other cols in the table, they wouldn't show up in
On Thu, Aug 6, 2009 at 3:14 PM, Michael Bayermike...@zzzcomputing.com wrote:
When populating objects through the ORM, I'd like to interpret all
NULL values fetched from VARCHAR2 / NVARCHAR2 columns in the database
as empty strings ('') instead of `None`s.
use a TypeDecorator that creates
On Aug 21, 2:17 am, King Simon-NFHD78 simon.k...@motorola.com
wrote:
darkblueB wrote:
ok, I have done this
I have an object def and a __table__ for all of the main actors in
my setup
I defined an __init__() for one of them
I use the declarative base
when the objects get created,
Hi Simon,
On Fri, 21 Aug 2009, King Simon-NFHD78 wrote:
Faheem Mitha wrote:
Thanks for the fast and helpful response. This looks like an artifact
of how I am creating the table. I wonder if this would still show up if
I explicitly specified the id. I could check this. Also, presumably
but I thought somehow that meta.reflect() would pick up on the fk
constraints ...
Class MetaData is part of the core api that ORM builds on top of, and
meta.reflect() does pick up the fk constraints from the database.
Think of it this way, MetaData, whether reflected or declared in its
Heyho!
Instead of creating changeby / changed fields on all my tables, I'm
planning to write some model classes where changes would be recorded in a
separate audit trail table (the obvious benefit beyond not requiring the
additional fields is that I can preserve the history as far back as I
You might want to start here
http://www.sqlalchemy.org/trac/wiki/UsageRecipes/LogVersions
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To post to this group, send email to
Ants Aasma ants.aa...@gmail.com writes:
This is something that could be improved in SQLAlchemy, but as a
workaround you can use the compiler extension to create the support
yourself. Here's some example code. It uses some private internals
from SQLAlchemy so you need to keep an eye on it
Adrian von Bidder avbid...@fortytwo.ch writes:
Ideas comments?
For what it's worth, I'd think that the best sort of audit would be
something done in the database itself, since it would audit any
changes whether done through any interface.
It depends on the database involved, but for example,
20 matches
Mail list logo
