OLEDB Command Parameter Type Is Different In Different Situations

Hi All,

I am using OLEDB Command transformation in a data flow to update the table. Find the columns for the table EMP as follows.


EmpID - int
EmpName - varchar(40)
EmpSal - float
Status - varchar(20)

I am using the following command to update the table.


Update EMP set status = 'Disabled' where EmpID = ? and EmpSal = ?


when I use the above condition the type of the second parameter is taking as "double precision float" as the incoming input column type is "double - precision float". This is fine.


But when I use the following condition the type of the param is taking as different one.

Update EMP set status = 'Disabled' where EmpID = ? and ( EmpSal + ? ) = 0


It is taking as "four byte signed integer" even though the incoming input column type is "double precision float" thus I am loosing the precision and getting error when the limit exeeds. If I use 0.0 then it is taking as "Numeric" type.


If I use convert function like (EmpSal + ? ) = Convert(float, 0) then it is taking as "double precision float".
Update EMP set status = 'Disabled' where EmpID = ? and ( EmpSal + ? ) = Convert(float, 0)


My question is how it behaves in the above situation. Can any body clarify please?

Venkat.