OK then, your ansisql patch is committed. we'll go with the simpler method for now. we can do version checks later on to support conditions for using ROW_NUMBER...do we have a newer MS-SQL server to test that with ? as for the OFFSET stuff, I would rather not build in "row consumption" as an extension to a particular kind of query, particularly since that totally breaks in the case of subqueries that specify OFFSET, as is the case with an eager-loading mapper that selects with an "offset" or "limit" parameter. I vote for raising the exception right now, so that an application can choose to define whatever scheme it wants to deal with that, but none by default. On Feb 18, 2006, at 5:03 PM, Rick Morrision wrote: Well, that won't work in pre-2005 SQL server, which are bound to be the overwhelming majority for the next few years. Perhaps we can check @@DBVERSION and act appropriately depending on the version, or make the 2005 version a different module (anybody know if pymssql / FreeTDS even works with MSSQL 2005?) |
- [Sqlalchemy-users] LIMIT vs TOP Rick Morrision
- Re: [Sqlalchemy-users] LIMIT vs TOP Jonathan Ellis
- Re: [Sqlalchemy-users] LIMIT vs TOP Michael Bayer
- Re: [Sqlalchemy-users] LIMIT vs TOP Rick Morrision
- Re: [Sqlalchemy-users] LIMIT vs TOP Rick Morrision
- Re: [Sqlalchemy-users] LIMIT vs TOP Daniel Miller
- Re: [Sqlalchemy-users] LIMIT vs TOP Michael Bayer
- Re: [Sqlalchemy-users] LIMIT vs T... Rick Morrision
- Re: [Sqlalchemy-users] LIMIT ... Michael Bayer
- Re: [Sqlalchemy-users] LIMIT ... Rick Morrision
- Re: [Sqlalchemy-users] LIMIT ... Michael Twomey