Toad saves file in UTF-8 with incorrect BOM (byte order mark).
The corruption is being introduced by the editor being used, as the script need to be modified before loading it into the target database (the "Disable constraints" option in "Generate Data Script" does not work properly because it disables them for a table, imports the table, then re-enables them for the table, which bombs if the table has any references to tables that have not been imported yet).
The problem is not actually with the editor being used, though. It's still with the script generated by Toad. The file is written with the UTF-8 byte order marks EF BB BF written TWICE to the file; those should be written only once. The editor being used gets confused by the duplicate BOM, strips the duplicate, but doesn't adjust the file buffer properly. When it saves the modified file, it introduces the corruption at 0x4000 because that is the block size it works with.
Stripping out one of the UTF-8 BOMs solves the problem.
The "Generate Data Script" should write only one UTF-8 BOM to the file.
WORKAROUND:
None.
STATUS:
Issue fixed in version 6.0. The latest version of Toad for SQL Server can be downloaded here.
