It must be a bad day for me. I can't make out what you're trying to accomplish. But, with the proper indexes, an INSERT might not block a SELECT on a table.

What has to happen to calculate the new check value? Is it something that can be embedded in an UPDATE query?

-Mike

>Hi Michael,
>
>Code schema:
>BEGIN TRAN
>query on T1 for filtered data to get a reliable check value
>-- no insert new data of type filtered on T1 ( this change check value )
>Control check value
>Use Check value to compute a new value
>use Check value to get data from other table
>other check control
>IF OK
> COMMIT
>ELSE
> ROLLBACK
>
>
>No insert is strong rule, but only for filtered data.
>
>i want' split T1 on multiple Table because filterd data range is smallint.
>
>For gain concurrency i wan't lock INSERT for not filtered data.
>
>Fabio