>Hi Alan,
>
>-----
>Well, maybe, but if you want to use binary data in your table, it seems to me that using a binary index might not be so absurd.
>
>My english is bad, but example is clear ( you have try it ? ):

I understood no problem. I did try it. Using "==" gives correct results. Using "=" does not.

>Except special situation USE "General" collate with binary index is a error.
>
>I put another example:
>

>CLEAR
>Create Cursor test(F1 C(2))
>
>Index On F1  tag Num COLLATE "general"
>
>Insert into test(F1) values ('A ')
>Insert into test(F1) values ('A'+CHR(0))
>Insert into test(F1) values ('A'+CHR(1))
>Insert into test(F1) values ('A'+CHR(2))
>Insert into test(F1) values ('A'+CHR(3))
>
>COUNT FOR F1=='A'+CHR(0)
>? _TALLY && this return 4 and is correct
>

Now I'm not sure of your point. The result is correct, so using 'Collate General' seems to be the best way to index when you are using binary data. Are you still saying that "Collate General" is not sensible?

Alan