Thanks to Erland Sommarskog ,
now I know that my problem with SQL
(Incorrect Sort of Farsi Unicode in SQL Server 2000) can not be solved :[
-----------------------


As you know, SQL Server currently does not support a Farsi or Arabic
collation. It depends upon Windows Locale to support Arabic and Farsi.

There was a bug in Windows relating to sorting of Farsi which was fixed.
But we still have some issues relating to that when SQL Server comes
into play.

There are 2 more bugs listed related to the issue. One of them relates
to SQL Server 2000 and the other to Yukon.

1. Bug # SQL Server 8.0 359136 - SQL sorts Farsi characters incorrectly
(SQL Server 2000)

2. BUG# SQL Server 9.0 523203 - Add Farsi collation. Dependent on NLS
addition of Farsi (YUKON)

Both of them are still active and they might be fixed in Yukon. But the
current status is active.

SQL currently does not support a Farsi collation. Books Online tells
users to map the Farsi Windows locale to the Arabic SQL collation, but
given that Arabic sorting rules do not (and should not) sort Farsi data
correctly, what is needed to fix this in SQL is a new Farsi collation
added to SQL that encapsulates the fixed Windows 2000 SP1 Farsi sorting
behavior. I cannot say when this will be implemented.

The good news is that Microsoft is aware of it, but the bad news is that it may not get fixed soon.

What you can do are two things:

1) Open a case yourself with Microsoft and refer to the bug number. If you
have compelling business reasons enough your case may result in a fix.
2) Send a mail to sqlwish@microsoft.com and request that a Farsi collation
is added to SQL Server. I know that the SQL Server developers pay
attention to what comes into this mailbox.