There are a couple of ways around that- setting the grid's recordsource to "" before you rerun the query is one- another is to zap the cursor, and append in the new results from some temp results. Both will avoid losing the grids settings.

>But don't you have problems when you recreate the SQL cursor with the column information?
>
>
>
>>You can use a cursor created by SQL as the recordsource for a grid- I do it all the time.
>>
>>>Erik,
>>>
>>>How about you want to use the view as a RecordSource for a Grid?
>>>
>>>Just a thought
>>>
>>>Rex
>>>
>>>>>Dear friends,
>>>>>
>>>>>I will try to answer to all...
>>>>>
>>>>>My code for the view is:
>>>>>
>>>>>select az.codArm, az.descricao, aaz.lote, Sum(aaz.quant) ;
>>>>> from tabAtAz aaz inner join tabAz az ;
>>>>> On az.stamp = aaz.st_tabAz Where aaz.st_tabAt = ?st_Artigo ;
>>>>> GROUP BY az.codArm, az.descricao, aaz.lote ORDER BY az.CodArm, aaz.lote
>>>>>
>>>>
>>>>I'm not sure wshat is causing the problem- I've never run into it because I have never tried to create a view with a GROUP BY clause. If your view does not need to be updatable, why not just use a plain ol' SQL query?