I have a stored procedure that generates a CSV file that is mailed to an automated process handler using xp_sendmail. In SQL Server version 7/NT 4, the CSV file was generated in code page 1252, but after migrating to SQL Server 2000/Win2K it is created in Unicode. This is a problem as the remote process handler doesn't handle this.
I installed SQL server with the same collation sequence as the version 7 server for compatibility purposes (SQL_Latin1_General_CP1_CI_AS).
Is there a setting that I can use to control how data is written to a flat file? If not, how can I change the SQL to force it to use a single-byte character set?