[sqlalchemy] Re: how to build a model package?
On Sat, Dec 20, 2008 at 9:34 AM, Etienne Robillard robillard.etie...@gmail.com wrote: On Wed, 17 Dec 2008 02:48:08 -0600 Jorge Vargas jorge.var...@gmail.com wrote: Hello, I have been trying to figure out an alternative to the following setup. The problem I'm trying to solve is fixing TurboGears project structure but I'm sure this is useful for other projects. model --- __init__.py --- foo.py bar.py in __init__.py we have the initialization of SA, creating mapper/session,etc. then both foo and bar import from model the session metadata, now that we need to initialize (bind several things) SA before using it, the current solution is to provide a function (init_model) that TG itselfs calls when the whole thing is loaded, the side effect this has is that we need to import all classes at the end of the file, which is at best ugly, but really confusing. So my question is, how do you handle a big set of classes that live in their own modules but can be imported directly from the package? Hi Jorge, I'm not sure understanding your question correctly. Do you search for ways to improve your model package based on sqlalchemy.ext.declarative or using classic SA techniques? not really, ideally this shouldn't be tied to any syntax, currently that is all build in in model.__init__.py Either ways, I kinda like making a model package (model.py) free of all the database-related stuff. I prefer putting theses things in database.py, so that its possible to modify the database options independently from the model. I think you got your concept wrong, model.py is a module not a package, a package is a directory with a __init__.py + several modules. If it was just wording yes this is what we are doing now, but we are talking about big models (think 50+ classes) your model.py will become huge and potentially messy, and if you split it into several files you will find that you lose the ability to do from myapp.model import Foo my custom app then look like so: mycustomapp/ - __init__.py - model.py - database.py model.py: from sqlalchemy.schema import MetaData, Table from sqlalchemy.ext.declarative import declarative_base import some_database as db __all__ = ['Base', 'metadata', 'BlogPost'] Base = declarative_base() metadata = MetaData(db.engine) class BlogPost(Base): ... I suppose the declarative_base() and SA 0.5.x greatly simplifies how models gets mapped internally. You can then just import your model class normally: from mycustomapp.model import BlogPost or even: import mycustom.model as model HTH, erob --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: 0.5 release schedule
I just launched last week a site running on 0.5rc4, given the quality of SA releases you shouldn't fear it's not final tag. It's really stable the only thing that could stop you from deploying is a hard policy on what goes to prod, by your organization/company/etc. On Thu, Dec 18, 2008 at 7:54 PM, Michael Bayer mike...@zzzcomputing.com wrote: I have put a site into production using 0.5rc4, and its a fairly prominent site for a popular TV show.0.5.0 is mostly ready to go, I'm holding off on some ORM enhancements to concrete inheritance until 0.5.1 since I need to study them further and concrete inheritance is not a popular feature. So what we're waiting for is a set of enhancements and fixes to the MS-SQL dialect, which will create a change in how URLs work, and then we're good to go. On Dec 18, 2008, at 6:02 PM, David Gardner wrote: I realize this may not be a fair questions since SA is an open source project. With that caveat what is the expected release date for 0.5 final? I have been developing against the 0.5 release candidates, and will probably deploy my code (an internal web app) in the second week of January. If 0.5 is out by then I will install and deploy that on my production server. If 0.5 isn't released I'm wondering if I should deploy my code against 0.4.8 or if I should deploy with 0.5rc4? I am probably more comfortable deploying with 0.5rc4 since that is what I have been developing against, and I am not sure if I have done anything that wouldn't work in 0.4.8. -- David Gardner Pipeline Tools Programmer, Sid the Science Kid Jim Henson Creature Shop dgard...@creatureshop.com --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] psycopg2 and Python 2.6
Hi all, I'm trying to port to python 2.6 with SQLAlchemy over PostgrSQL and it seems that all is porting BUT the psycopg2 component? Is anyone aware of a working version of psycopg2 for Linx 64 arch? In general - does SQLAlchemy support 2.6? I was under the impression that it does. Is there another DB (not PostgerSQL) that does work? Thanks --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: MSSQL default_schema
I've got a few concerns with the just-committed get_default_schema_name() approach: 1) What about pre-2005 users? The query used in this patch to fetch the schema name won't work. There was not even a real hard concept of 'schema' pre-MSSQL-2005. 2) Prior to MSSQL 2005, MSSQL conflated user name and schema name in a fashion similar to that of Oracle. People use this to override some tables, because the server search for locating a table in Oracle and MSSQL is to check the user schema first, then look through a schema search path looking for a table matching the given identifier. So if you have a table named [mylogin.tablename] and there is also a table [public.tablename], if you issue SELECT * FROM tablename, you'll get the contents of [public.tablename] UNLESS you're logged in as 'mylogin', in which case you'll get the contents of [mylogin.tablename]. This is IMO a kind of questionable practice, but there are existing uses like this out there, and making the default schema always explicit breaks this. Note this is not only a pre-MSSQL2005 issue, I think 2005+ will still recognize user style schemas. 3) Isn't Micheal's concern below still valid? just FTR, the current expected behavior of default schemas is that if your tables are known to exist in the default schema configured on the database connection, you leave the schema attribute on Table blank. otherwise, you set it. this is actually a convention that we've had to choose. if the two conventions are mixed, table reflection gets confused over whether to look for/place tables with the key tablename or schemaname.tablename into the MetaData object. While we could try building an enhanced key object which hashes both tablename and schema.tablename as the same key depending on the default schema of the DB, this requires a live connection to the database in order to function, and as we all know a MetaData can be bound to a DB later on...so for now we have the convention. --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: MSSQL default_schema
Rick, On Tue, Dec 23, 2008 at 10:32 AM, Rick Morrison rickmorri...@gmail.comwrote: I've got a few concerns with the just-committed get_default_schema_name() approach: 1) What about pre-2005 users? The query used in this patch to fetch the schema name won't work. There was not even a real hard concept of 'schema' pre-MSSQL-2005. I looked at that. It looks like how it's constructed it will only affect users of MSSQL 2005. MSSQL 2000 and before should get the same behavior they had previously due to the try / except. We can also put in a version check if preferable. Michael --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] ORM relations...
Over the last couple days I have been having an issue with relations/ imports. I have a class called queue where I define all my columns. Within queue I have a function that references a class filequeue so I need to 'from filequeues import filequeue'. The queue class itself has relations with other classes/tables. My problem is that within filequeue, I need to 'from queues import queue' in order to ForeignKey (queue.queueid). Well now queue imports filequeue and filequeue imports queue causing an issue. I tried to ForeignKey ('QUEUES.queueid') but SQLAlch says it cannot find the tables. Any ideas? --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: psycopg2 and Python 2.6
we're good for 2.6 and in particular the 0.5.0 trunk (post rc4) is tested on 2.6, and even runs fairly well with the -3 flag. On Dec 23, 2008, at 8:04 AM, vctr...@gmail.com wrote: Hi all, I'm trying to port to python 2.6 with SQLAlchemy over PostgrSQL and it seems that all is porting BUT the psycopg2 component? Is anyone aware of a working version of psycopg2 for Linx 64 arch? In general - does SQLAlchemy support 2.6? I was under the impression that it does. Is there another DB (not PostgerSQL) that does work? Thanks --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: ORM relations...
ForeignKey accepts strings to remove the inconvenience of handling circular imports. Both dependent tables will need to be present within the shared MetaData object at the point at which the Table object is first used, though, so another module that imports both queue and filequeue would be best. On Dec 23, 2008, at 10:48 AM, justmike2...@gmail.com wrote: Over the last couple days I have been having an issue with relations/ imports. I have a class called queue where I define all my columns. Within queue I have a function that references a class filequeue so I need to 'from filequeues import filequeue'. The queue class itself has relations with other classes/tables. My problem is that within filequeue, I need to 'from queues import queue' in order to ForeignKey (queue.queueid). Well now queue imports filequeue and filequeue imports queue causing an issue. I tried to ForeignKey ('QUEUES.queueid') but SQLAlch says it cannot find the tables. Any ideas? --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: ORM relations...
I am not sure based on the project if I can have a module that imports both of them. How can I use the string based ForeignKey? I may be doing it in correctly because it says the table does not exist that I try to reference. On Tue, Dec 23, 2008 at 12:18 PM, Michael Bayer mike...@zzzcomputing.comwrote: ForeignKey accepts strings to remove the inconvenience of handling circular imports. Both dependent tables will need to be present within the shared MetaData object at the point at which the Table object is first used, though, so another module that imports both queue and filequeue would be best. On Dec 23, 2008, at 10:48 AM, justmike2...@gmail.com wrote: Over the last couple days I have been having an issue with relations/ imports. I have a class called queue where I define all my columns. Within queue I have a function that references a class filequeue so I need to 'from filequeues import filequeue'. The queue class itself has relations with other classes/tables. My problem is that within filequeue, I need to 'from queues import queue' in order to ForeignKey (queue.queueid). Well now queue imports filequeue and filequeue imports queue causing an issue. I tried to ForeignKey ('QUEUES.queueid') but SQLAlch says it cannot find the tables. Any ideas? --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: ORM relations...
thats because both Table objects need to be present in the MetaData when used. Ultimately your application can't run unless both queue and filequeue have been imported in some way. On Dec 23, 2008, at 1:21 PM, Michael Mileusnich wrote: I am not sure based on the project if I can have a module that imports both of them. How can I use the string based ForeignKey? I may be doing it in correctly because it says the table does not exist that I try to reference. On Tue, Dec 23, 2008 at 12:18 PM, Michael Bayer mike...@zzzcomputing.com wrote: ForeignKey accepts strings to remove the inconvenience of handling circular imports. Both dependent tables will need to be present within the shared MetaData object at the point at which the Table object is first used, though, so another module that imports both queue and filequeue would be best. On Dec 23, 2008, at 10:48 AM, justmike2...@gmail.com wrote: Over the last couple days I have been having an issue with relations/ imports. I have a class called queue where I define all my columns. Within queue I have a function that references a class filequeue so I need to 'from filequeues import filequeue'. The queue class itself has relations with other classes/tables. My problem is that within filequeue, I need to 'from queues import queue' in order to ForeignKey (queue.queueid). Well now queue imports filequeue and filequeue imports queue causing an issue. I tried to ForeignKey ('QUEUES.queueid') but SQLAlch says it cannot find the tables. Any ideas? --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: ORM relations...
I have a py file that creates my tables. It does MetaData = queue.metadata. metadata.create_all(). That's all I do with MetaData. Is there additional table setup I need to be doing? On Tue, Dec 23, 2008 at 12:24 PM, Michael Bayer mike...@zzzcomputing.comwrote: thats because both Table objects need to be present in the MetaData when used. Ultimately your application can't run unless both queue and filequeue have been imported in some way. On Dec 23, 2008, at 1:21 PM, Michael Mileusnich wrote: I am not sure based on the project if I can have a module that imports both of them. How can I use the string based ForeignKey? I may be doing it in correctly because it says the table does not exist that I try to reference. On Tue, Dec 23, 2008 at 12:18 PM, Michael Bayer mike...@zzzcomputing.comwrote: ForeignKey accepts strings to remove the inconvenience of handling circular imports. Both dependent tables will need to be present within the shared MetaData object at the point at which the Table object is first used, though, so another module that imports both queue and filequeue would be best. On Dec 23, 2008, at 10:48 AM, justmike2...@gmail.com wrote: Over the last couple days I have been having an issue with relations/ imports. I have a class called queue where I define all my columns. Within queue I have a function that references a class filequeue so I need to 'from filequeues import filequeue'. The queue class itself has relations with other classes/tables. My problem is that within filequeue, I need to 'from queues import queue' in order to ForeignKey (queue.queueid). Well now queue imports filequeue and filequeue imports queue causing an issue. I tried to ForeignKey ('QUEUES.queueid') but SQLAlch says it cannot find the tables. Any ideas? --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: ORM relations...
On Dec 23, 2008, at 1:29 PM, Michael Mileusnich wrote: I have a py file that creates my tables. It does MetaData = queue.metadata. metadata.create_all(). That's all I do with MetaData. Is there additional table setup I need to be doing? yeah have it do import filequeue; import queue; queue.metadata.create_all(). or alternatively have queue import filequeue as well. --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: MSSQL default_schema
Im curious, is the MSSQL dialect rendering tables as schemaname.tablename in all cases ? No, I don't think so: the module uses non-overridden calls to compiler.IdentifierPreparer.format_table() and format_column(). So then the only usage of the get_default_schema_name() is for table existence checks, and to fetch schema metadata for table reflection? Then I'll withdraw my concerns, that's no big deal. --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: MSSQL default_schema
On Dec 23, 2008, at 2:15 PM, Rick Morrison wrote: Im curious, is the MSSQL dialect rendering tables as schemaname.tablename in all cases ? No, I don't think so: the module uses non-overridden calls to compiler.IdentifierPreparer.format_table() and format_column(). So then the only usage of the get_default_schema_name() is for table existence checks, and to fetch schema metadata for table reflection? Then I'll withdraw my concerns, that's no big deal. so if we can confirm this ! .socloseto...release --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: MSSQL default_schema
Michael Bayer wrote: so if we can confirm this ! .socloseto...release From grepping: For the Oracle dialect get_default_schema_name is used in reflecttable, has_table and has_sequence. has_table, in turn, is used by the SchemaGenerator and SchemaDropper. MSSQL uses it for reflecttable and has_table and others do the same or less. Outside of the dialects, get_default_schema_name is used for Engine.table_names -Randall --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: cascade
ondelete and onupdate are keyword arguments within ForeignKey() itself, and they come out as part of the DDL when you call metadata.create_all(). On Dec 23, 2008, at 8:42 PM, justmike2...@gmail.com wrote: I seem to be running into a problem with implementing cascades in sqlalch. Next to my ForeignKey() I have ondelete and onupdate both set to CASCADE and in my mapper I have this: mapper(queue, queue_table, properties=dict(parent=relation(queue, backref='JOBS'))) However the cascade is not being listed in MSSQL. --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: MSSQL default_schema
On Tue, Dec 23, 2008 at 3:34 PM, Randall Smith rand...@tnr.cc wrote: Michael Bayer wrote: so if we can confirm this ! .socloseto...release From grepping: For the Oracle dialect get_default_schema_name is used in reflecttable, has_table and has_sequence. has_table, in turn, is used by the SchemaGenerator and SchemaDropper. MSSQL uses it for reflecttable and has_table and others do the same or less. Outside of the dialects, get_default_schema_name is used for Engine.table_names If you guys tell me how, I could run the unit test for sqlalchemy on 2005, and 2000 sql server. Lucas --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: cascade
Yea I do the ondelete and onupdate in ForeignKey, creates tables and all but does not do the cascade or at least not in MSSQL. On Tue, Dec 23, 2008 at 7:42 PM, justmike2...@gmail.com justmike2...@gmail.com wrote: I seem to be running into a problem with implementing cascades in sqlalch. Next to my ForeignKey() I have ondelete and onupdate both set to CASCADE and in my mapper I have this: mapper(queue, queue_table, properties=dict(parent=relation(queue, backref='JOBS'))) However the cascade is not being listed in MSSQL. --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: cascade
generates the ON DELETE CASCADE in the DDL, versus the cascade does not work, are two different issues. Which are you experiencing, and if the latter, how is the issue manifesting itself ? Can you produce sample code which reproduces your issue ? On Dec 23, 2008, at 10:19 PM, Michael Mileusnich wrote: Yea I do the ondelete and onupdate in ForeignKey, creates tables and all but does not do the cascade or at least not in MSSQL. On Tue, Dec 23, 2008 at 7:42 PM, justmike2...@gmail.com justmike2...@gmail.com wrote: I seem to be running into a problem with implementing cascades in sqlalch. Next to my ForeignKey() I have ondelete and onupdate both set to CASCADE and in my mapper I have this: mapper(queue, queue_table, properties=dict(parent=relation(queue, backref='JOBS'))) However the cascade is not being listed in MSSQL. --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: cascade
code: queue_table = Table( 'QUEUES', metadata, Column('QUEUEID', Integer, autoincrement=True, primary_key=True), Column('JOBNUM', String(48), ForeignKey('JOBS.JOBNUM', onupdate=CASCADE, ondelete=CASCADE)), Column('SCHEDID', String(48), ForeignKey('SCHEDULES.SCHEDID')), Column('DATE', DateTime), Column('STATUS', Integer)) # 0 sched 1 Ignore 2 Evalling 3 Running 4 Unsched 5 Error mapper(queue, queue_table, properties=dict(parent=relation(queue, backref='job'))) In MSSQL the cascade entries for those tables are set to none and my app also generate this: many-to-many relation, 'secondaryjoin' is needed as well. % (self)) sqlalchemy.exc.ArgumentError: Could not determine join condition between parent/ child tables on relation queue.parent. Specify a 'primaryjoin' expression. If this is a many-to-many relation, 'secondaryjoin' is needed as well. On Tue, Dec 23, 2008 at 10:52 PM, Michael Bayer mike...@zzzcomputing.comwrote: generates the ON DELETE CASCADE in the DDL, versus the cascade does not work, are two different issues. Which are you experiencing, and if the latter, how is the issue manifesting itself ? Can you produce sample code which reproduces your issue ? On Dec 23, 2008, at 10:19 PM, Michael Mileusnich wrote: Yea I do the ondelete and onupdate in ForeignKey, creates tables and all but does not do the cascade or at least not in MSSQL. On Tue, Dec 23, 2008 at 7:42 PM, justmike2...@gmail.com justmike2...@gmail.com wrote: I seem to be running into a problem with implementing cascades in sqlalch. Next to my ForeignKey() I have ondelete and onupdate both set to CASCADE and in my mapper I have this: mapper(queue, queue_table, properties=dict(parent=relation(queue, backref='JOBS'))) However the cascade is not being listed in MSSQL. --~--~-~--~~~---~--~~ 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 sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---