o kas dokumentacija skaitys? :)
When the query specified returns multiple results, the result set for each 
row returning query is placed in a separate table. Additional result sets 
are named by appending integral values to the specified table name (for 
example, "Table", "Table1", "Table2", and so on). Because no table is 
created for a query that does not return rows, if you process an insert 
query followed by a select query, the table created for the select query is 
named "Table" because it is the first table created. Applications using 
column and table names should ensure that conflicts with these naming 
patterns does not occur.

When the SELECT statement used to populate the DataSet returns multiple 
results, such as batch SQL statements, if one of the results contains an 
error, all subsequent results are skipped and are not added to the DataSet.


"Jornada Del Muerto" <ask@me.email> wrote in message 
news:hk9a31$rns$1@trimpas.omnitel.net...
> Sveiki,
>
>    Tokia briedova ideja kilo, pvz turiu MSSQL ir as galiu proceduroje 
> daryti:
>
>    CREATE PROCEDURE MyProc AS
>
>    Select * From lenta1 where type=1
>    Select * From lenta1 where type=10
>    Select * From lenta2
>    Select * From lenta3
>
>    Tada jei kokiam nors manageryje paleisit tokia proc tai kaip visi zino 
> grazins kelis rezultatus, o ar eitu taip kad ir i DataSet'a paimt per 
> ADO.NET
>
>
>    Tipo daugmaz standartinis kreipimasis i SQL ir DataSet.Load, yra siaip 
> toks metodas bet jis tada klausai lentu vardu, bet as gal 10 selectu 
> skirtingu is tos pacios darau...
>
>    Dar yra DataAdapter.Fill bet as negaliu gauti is IDbConnection jo 
> IDataAdapter, nes dirbu interface'u lygyje...
>
>    Gal kas bande? pavyko nepavyko? ar siaip turi nuomuone ?
>