MSSQL dialect needs to filter index column names

[Former user]

The MSSQL dialect's base::get_indexes() method needs to watch out for column names that end with "(-)".

http://msdn.microsoft.com/en-us/library/ms188771.aspx

A descending indexed column will be listed in the result set with a minus sign (-) following its name; an ascending indexed column, the default, will be listed by its name alone.

Fix:

Patch dialects/mssql/base.py::get_indexes to find and remove that postfix.

[Former user]

patch to fix column names in indexes

[Mike Bayer]

can we replace the various lambda/strip stuff just with a simple regex to clean that up.

test coverage in test/dialect/test_mssql.py is highly recommended here as well.

[Former user]

converted column_names extraction to use a regex

[Former user]

regex.patch as unified diff

[Former user]

regex.patch failed one hunk for me and added the other hunk in the wrong place.

I attached fix_descending_column_names.patch which is the above patch as a unified diff.

It applies cleanly from the top of the trunk as of f72378137a855098c4f0c1c830e3b244634cbd84

hopefully it can land soon...

[Mike Bayer]

ideally we'd have a test for this but its good to go in abdf3a22ccb23f2cfc7e5cf8b815c62f925d7b0d

[Mike Bayer]

Removing milestone: 0.6.0 (automated comment)