Olivier,

Thanks for your compliment. I'm glad you like this article.

In addition to -c -q -S switches in my code, try adding -w in both bcp scripts. -w is the unicode switch. With that switch, BCP will perform the bulk copy operation using Unicode characters. It uses nchar as the storage type. Hopefully that will take care of accented letters like é, è, ê, û. According to Books on Line, -w will not work for SQL Server 6.5.

All systems I have are in American English collation/code page, therefore I couldn't test that for you. I am curious to know if that solves your problem.

Haidong