PRB: #deleted seen when SQL Server 7.0 tables or SQL Server 6.5 tables that are linked to Access 97 or Access 2000 are opened

If Microsoft Access 97 or Microsoft Access 2000 is used to link SQL Server 6.5 or 7.0 tables that have numeric and text field columns, #deleted is seen in alternate rows if the SQL Server database is created by using non default code pages and Sort Order. This behavior can be reproduced when code pages are cp850 or cp437, and the sort order is binary, case insensitive.

This error is only seen with the Microsoft ODBC SQL Server Driver version 3.70.0623.

To work around this problem, clear the Perform translation for character data option when configuring the data source name (DSN).

This problem does not exist with previous or later versions of the Microsoft ODBC SQL Server Driver.

Additionally, using the Microsoft ODBC SQL Server Driver version 3.70.0623 with a pass-through query does not generate the #delete error.

Steps to reproduce the behavior

/* From the ODBC Administrator version 3.510.3711 (MDAC 2.1 SP1(GA)) */

1. Create a system DSN that uses SQL Server 7.0, using the default settings, and connect to the pubs database.
2. Using Microsoft Access, link the Employee and Discounts tables through the DSN created in step 1. Open the linked tables and you will see #deleted in alternate rows.
3. Run a SQL specific pass-through query using the same DSN and then run either of the following SQL statements:

      Select * from employee
      Select * from discounts
   						

   The data returns correctly unlike the data in the linked tables.
4. Close Microsoft Access 97 or 2000. Open the ODBC Administrator and re-configure the system DSN so that the Perform translation for character data option check box is clear (the check box is selected by default).
5. Open Microsoft Access 97 or 2000 again, and then connect to these linked tables and the data returns correctly.
6. Run a SQL specific pass-through query using the re-configured DSN, and then run either of the following SQL statements :

      Select * from employee
      Select * from discounts
   					

   Again, the data returns correctly.