Sure and thanks for your attention

The result I want is something like the tblSales. First I calculate the DataBeginSales from the table historicoproduto.for each product.Next for each product I calculate the sum of sales from the table romaneio and update the cursor.

I cannot figure out how to do that without using cursors, in a set-based way for better performance.

declare @DateINi smalldatetime
declare @TotSales Numeric(10,2)

set @DateINi='20100218'

select @dateini=dateadd(hh,(select datepart(hh,getdate())),cast(convert(CHAR(11),@dateini,113) as datetime))
select @dateini=dateadd(mi,(select datepart(mi,getdate())),@dateini)
select @dateini=dateadd(ss,(select datepart(ss,getdate())),@dateini)

print @dateini

select @TotSales=(select sum(valor) from romaneio r  where tipo='VC' and caddate>=@dateini)

declare  SSales cursor for
	select e.id,max(caddate) as DataBeginSales from historicoproduto h right join estoquefisico e
on h.prodid=e.id
  where ativo=1 and ( tipo IS NULL or tipo<>'Posição'
 or (tipo='Posição' and (h.historico like 'Colocação em Venda%' or h.historico like '%->Venda%'))) group by e.id

DECLARE @tblSales TABLE (ProdID Char(10),Data smalldatetime,Sales Numeric(10,2))
DECLARE @ProdID char(10)
DECLARE @MaxSales smalldatetime
DECLARE @Sales Numeric(10,2)

OPEN SSales

 

--Atribuindo valores do select nas variáveis

FETCH NEXT FROM SSales INTO @ProdID, @MaxSales


--Iniciando laço
WHILE @@FETCH_STATUS = 0

BEGIN
   SET @Sales=@TotSales
   IF dbo.udf_TempoDecorridoSegundos(@MaxSales,getdate())<180000
		SET @Sales=(select sum(valor) from romaneio r  where tipo='VC' and caddate>=@maxSales)
   
	INSERT INTO @tblSales values(@ProdID, @MaxSales,@Sales)


--Próxima linha do cursor

    FETCH NEXT FROM SSales INTO @ProdID, @MaxSales

END

 

--Fechando e desalocando cursor

CLOSE SSales

DEALLOCATE SSales

select * from @tblSales