Q1 – In the code example below runtime converts a value between two
incompatible types:
<SelectParameters>
<asp:Parameter Name="City" Type="Int32" />
</SelectParameters>
protected void SqlDataSource2_Selecting(object sender,
SqlDataSourceSelectingEventArgs e)
{e.Command.Parameters["@City"].Value = "100";}
Exception I got:
“Conversion failed when converting the nvarchar value 'Seattle' to
data type int.”
A) The above exception suggests that runtime did manage to convert a
value of type String into a value of type Int32, and thus the
exception happened on SqlServer?!
B) Since String and Int32 are incompatible types, why did runtime
perform conversion from String to Int32 in the first place?
Doesn’t the fact that we’re dealing with incompatible type make
runtime “realize” that app most likely has a bug, similary to the way
compiler “realizes” that it’s dealing ( in the code below ) with two
incompatible types:
String s = ”something”;
int i = (int)s; //error
Q2:
A)
public void GetEmployee( int EmployeeID );
<asp:ObjectDataSource SelectMethod=”GetEmployee” …>
<SelectParameters>
<asp:ControlParameter Name = ”EmployeeID” ...>
</SelectParameters>
If for whatever reason EmployeeID parameter is NULL, ObjectDataSource
will convert Null to zero and passed it as argument to GetEmployee()
method.
Why does runtime make such a conversion? Wouldn't throwing an
exception made more sense?
B) “Use the ConvertEmptyStringToNull property to specify whether an
empty string value is automatically converted to null when the data
field is updated in the data source.”
I don’t quite understand the usefulness of this property. Why would
empty string indicate that we want null to be inserted into source’s
data field? I’d assume that this data field is of type String? Then
why not also have ConvertZeroInt32ToNull etc?
Q3:
<asp:SqlDataSource ID="sourceEmployees" runat="server"
ProviderName="System.Data.SqlClient"
ConnectionString="<%$ ConnectionStrings:Northwind %>"
SelectCommand="SELECT EmployeeID, FirstName,LastName,
Title, City FROM Employees WHERE City=@City">
<SelectParameters>
<asp:Parameter Name="City"/>
</SelectParameters>
</asp:SqlDataSource>
A) I assume that when you don’t specify of which type Parameter
instance “City” is, it is automatically of type Object, which means it
can later be assigned value of any type. Thus if “City” is later ( say
inside SqlDataSource2_Selecting() event handler ) assigned a value of
a wrong type, this wrong assignment will only be detected on Sql
server, and not before ( of course Sql server will report that error
back to web server )?
B) If we create a SqlParameter instance of type NVarChar(20) and want
to pass this parameter to a stored procedure, will Ado.net pass to a
stored procedure just the value of this parameter, or will it also
somehow inform the procedure the exact type of this parameter ( which
is NVarChar(20))?
thanx
Best Answer
This is a SQL Server error. You've hit a datatype precedence issue.
The clue is here:
The City column is nvarchar The parameter is explicitly set to int
Adding it all together, you get:
By the rules of datatype precedemce, the database engine tries to change all the values for City to "int". And fails, of course.
For your points though:
Overall, SQL Server is behaving exactly as advertised. In this case, you simply are not giving enough information.
Edit: Stored proc usage
You tell SQL what the data type is when you add the parameter to the parameters collection. You define (and create) the stored proc to have the nvarchar parameter.