[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-21 Thread Michael Bayer
I put up ticket #487 for this. theres a whole batch of sql-related tickets im going to try to line up for a marathon session at some future date in the next few weeks. On Feb 21, 2007, at 10:25 AM, Troy wrote: Mike, Based on your excellent feedback and trying a couple different

[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-19 Thread Troy
It's not about case in the resultproxy, it's about case-insensitive for server-side compare, such as in the where clause and when ordering results. comparisonslike literal text injected into the SQL? why not use bind parameters ?there are cases where literal text should not be

[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-19 Thread Michael Bayer
On Feb 19, 2007, at 4:52 AM, Troy wrote: It's not just the literal text or the bind param. It is how the server compares character based data. In MySQL, MS-SQL, Sybase -- case does not matter. In Postgres, Oracle and DB2 it does. DB2 and Oracle (since version 10 I think) have some

[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-19 Thread Troy
OK totally different issue. MySQL does have a case-sensitivity setting using COLLATE. not sure what MS-SQL has and we dont yet have sybase support. we've had people report the whole lower() issue in the past and im not sure that should be automatic within SA.You can just explicitly

[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-19 Thread Michael Bayer
On Feb 19, 2:32 pm, Troy [EMAIL PROTECTED] wrote: Even though MySQL allows the setting of COLLATE, it does not support functional indexes, so if your code explicitly calls lower you technically now have code that will work for both MySQL and Postgres, but MySQL is going to take a big

[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-18 Thread Troy
i suppose this is not terribly hard since it would just mean some extra argument passed along. of course the question is *why* would you need this. if its some kind of case-insensitive resultproxy issue, id rather do something on that end. It's not about case in the resultproxy, it's about

[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-18 Thread Michael Bayer
On Feb 18, 2007, at 6:37 PM, Troy wrote: i suppose this is not terribly hard since it would just mean some extra argument passed along. of course the question is *why* would you need this. if its some kind of case-insensitive resultproxy issue, id rather do something on that end. It's

[sqlalchemy] Re: Differentiate ANSIIdentifierPreparer.format_column select clause from where clause??

2007-02-16 Thread Michael Bayer
On Feb 16, 2007, at 3:18 AM, Troy wrote: Is there a way in the format_column method of ANSIIdentifierPreparer to determine if the column is part of the select clause, where clause, order clause, etc? What I'm attempting to do is override the default postgres dialect to format string