It wouldn't work if you have more than 10 duplicate values or duplicates begin in the previous batch.
>Hi Sergey, I was think this, but in big tables the query is too slow, because the combination of this, not exist in any index expression. I think that the best option is send a cursor with unique field that be the primary key field, and the component convert this XML to cursor and compare this in subquery, for exclude this records of the next or previus query.
>
>Something as this:
>
>SELECT TOP 10 field1,field2,field3 FROM datos ;
> WHERE field3 NOT IN ;
> (SELECT field3 FROM curtemp) ;
> ORDER BY field1
>
>Do I explain well?
>
>Tks for your sugerencies
--sb--