As it happens, this works on the Sybase dialect without fixing the
quoting at all! Apparently SQL such as this is happily accepted by
Sybase:
SELECT [fdcommon.dbo].organization.org_id,
[fdcommon.dbo].organization.abbrev
FROM [fdcommon.dbo].organization JOIN [fdcommon.dbo].org_type
ON [fdcommon.dbo].org_type.org_type_id =
[fdcommon.dbo].organization.org_type
I resorted to some brute-force list operations rather than regular
expressions to parse out the component names (see diff below). I will
fix the quoting shortly (within the next day or so) and submit a
single diff.
thanks,
pjjH
Index: schema.py
===================================================================
--- schema.py (revision 5816)
+++ schema.py (working copy)
@@ -876,17 +876,22 @@
raise exc.ArgumentError(
"Parent column '%s' does not descend from a "
"table-attached Column" % str(self.parent))
- m = re.match(r"^(.+?)(?:\.(.+?))?(?:\.(.+?))?$",
self._colspec,
- re.UNICODE)
+ m = self._colspec.split('.')
if m is None:
raise exc.ArgumentError(
"Invalid foreign key column specification: %s" %
self._colspec)
- if m.group(3) is None:
- (tname, colname) = m.group(1, 2)
+
+ m.reverse()
+ (colname, tname) = m[0:2]
+
+ if m[2] is None:
schema = None
else:
- (schema, tname, colname) = m.group(1, 2, 3)
+ m1 = m[2:]
+ m1.reverse()
+ schema = '.'.join(m1)
+
On Mar 5, 7:21 pm, "[email protected]" <[email protected]>
wrote:
> OK. If it might be as easy as that, I will have a go and see how well
> it works.
>
> pjjH
>
> On Mar 5, 4:31 pm, "Michael Bayer" <[email protected]> wrote:
>
> > [email protected] wrote:
>
> > > Sybase (and SQL Server) support cross-database JOINs (Sybase even
> > > supports cross-database foreign-key constraints). There are four
> > > components to an object identifier:
>
> > > 1 = Object name
> > > 2 = Schema name
> > > 3 = Database name
> > > 4 = Server name
>
> > > the dataserver, database and schema are assumed for one-part
> > > identifiers (e.g. 'foo'), dataserver and database assumed for two-part
> > > identifiers (e.g. 'dbo.foo') and, finally, dataserver assumed for
> > > three-part identifiers ('production.dbo.foo')
>
> > > e.g. SELECT foo.* FROM BANANA.production.dbo.foo
> > > SELECT foo.* FROM production.dbo.foo
> > > SELECT foo.* FROM production..foo -- same as the previous query
> > > if the callers default schema is dbo
> > > SELECT foo.* FROM dbo.foo
> > > SELECT foo.* FROM foo
> > > SELECT foo.* FROM ..foo
> > > I am not so interested in supporting four-part identifiers in SA but I
> > > would like to figure out how to support three-part identifiers as very
> > > many of our databases have cross-database references.
>
> > > One natural (to me!) way of doing this is to add a 'database' property
> > > to the Table and ForeignKeyConstraint schema items and have the Sybase/
> > > SQL Server dialects always emit fully-qualified three-part identifiers
> > > for table names.
>
> > we have an element on Table called "schema". I had in mind that schema
> > should accept dotted names, so SQLA generally doesn't need to get
> > involved. The only change needed is to the IdentifierPreparer, such that
> > when quote_schema is called, it separates the name along the dot first so
> > that it can quote each token separately.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~----------~----~----~----~------~----~------~--~---
