[c#] in a "using" block is a SqlConnection closed on return or exception?

First question:
Say I have

using (SqlConnection connection = new SqlConnection(connectionString))
{
    connection.Open();

    string storedProc = "GetData";
    SqlCommand command = new SqlCommand(storedProc, connection);
    command.CommandType = CommandType.StoredProcedure;
    command.Parameters.Add(new SqlParameter("@EmployeeID", employeeID));

    return (byte[])command.ExecuteScalar();
}

Does the connection get closed? Because technically we never get to the last } as we return before it.

Second question:
This time I have:

try
{
    using (SqlConnection connection = new SqlConnection(connectionString))
    {
        int employeeID = findEmployeeID();

        connection.Open();
        SqlCommand command = new SqlCommand("UpdateEmployeeTable", connection);
        command.CommandType = CommandType.StoredProcedure;
        command.Parameters.Add(new SqlParameter("@EmployeeID", employeeID));
        command.CommandTimeout = 5;

        command.ExecuteNonQuery();
    }
}
catch (Exception) { /*Handle error*/ }

Now, say somewhere in the try we get an error and it gets caught. Does the connection still get closed? Because again, we skip the rest of the code in the try and go directly to the catch statement.

Am I thinking too linearly in how using works? ie Does Dispose() simply get called when we leave the using scope?

This question is related to c# using sqlconnection

The answer is


Here is my Template. Everything you need to select data from an SQL server. Connection is closed and disposed and errors in connection and execution are caught.

string connString = System.Configuration.ConfigurationManager.ConnectionStrings["CompanyServer"].ConnectionString;
string selectStatement = @"
    SELECT TOP 1 Person
    FROM CorporateOffice
    WHERE HeadUpAss = 1 AND Title LIKE 'C-Level%'
    ORDER BY IntelligenceQuotient DESC
";
using (SqlConnection conn = new SqlConnection(connString))
{
    using (SqlCommand comm = new SqlCommand(selectStatement, conn))
    {
        try
        {
            conn.Open();
            using (SqlDataReader dr = comm.ExecuteReader())
            {
                if (dr.HasRows)
                {
                    while (dr.Read())
                    {
                        Console.WriteLine(dr["Person"].ToString());
                    }
                }
                else Console.WriteLine("No C-Level with Head Up Ass Found!? (Very Odd)");
            }
        }
        catch (Exception e) { Console.WriteLine("Error: " + e.Message); }
        if (conn.State == System.Data.ConnectionState.Open) conn.Close();
    }
}

* Revised: 2015-11-09 *
As suggested by NickG; If too many braces are annoying you, format like this...

using (SqlConnection conn = new SqlConnection(connString))
   using (SqlCommand comm = new SqlCommand(selectStatement, conn))
   {
      try
      {
         conn.Open();
         using (SqlDataReader dr = comm.ExecuteReader())
            if (dr.HasRows)
               while (dr.Read()) Console.WriteLine(dr["Person"].ToString());
            else Console.WriteLine("No C-Level with Head Up Ass Found!? (Very Odd)");
      }
      catch (Exception e) { Console.WriteLine("Error: " + e.Message); }
      if (conn.State == System.Data.ConnectionState.Open) conn.Close();
   }

Then again, if you work for EA or DayBreak games, you can just forgo any line-breaks as well because those are just for people who have to come back and look at your code later and who really cares? Am I right? I mean 1 line instead of 23 means I'm a better programmer, right?

using (SqlConnection conn = new SqlConnection(connString)) using (SqlCommand comm = new SqlCommand(selectStatement, conn)) { try { conn.Open(); using (SqlDataReader dr = comm.ExecuteReader()) if (dr.HasRows) while (dr.Read()) Console.WriteLine(dr["Person"].ToString()); else Console.WriteLine("No C-Level with Head Up Ass Found!? (Very Odd)"); } catch (Exception e) { Console.WriteLine("Error: " + e.Message); } if (conn.State == System.Data.ConnectionState.Open) conn.Close(); }

Phew... OK. I got that out of my system and am done amusing myself for a while. Carry on.


In your first example, the C# compiler will actually translate the using statement to the following:

SqlConnection connection = new SqlConnection(connectionString));

try
{
    connection.Open();

    string storedProc = "GetData";
    SqlCommand command = new SqlCommand(storedProc, connection);
    command.CommandType = CommandType.StoredProcedure;
    command.Parameters.Add(new SqlParameter("@EmployeeID", employeeID));

    return (byte[])command.ExecuteScalar();
}
finally
{
    connection.Dispose();
}

Finally statements will always get called before a function returns and so the connection will be always closed/disposed.

So, in your second example the code will be compiled to the following:

try
{
    try
    {
        connection.Open();

        string storedProc = "GetData";
        SqlCommand command = new SqlCommand(storedProc, connection);
        command.CommandType = CommandType.StoredProcedure;
        command.Parameters.Add(new SqlParameter("@EmployeeID", employeeID));

        return (byte[])command.ExecuteScalar();
    }
    finally
    {
        connection.Dispose();
    }
}
catch (Exception)
{
}

The exception will be caught in the finally statement and the connection closed. The exception will not be seen by the outer catch clause.


I wrote two using statements inside a try/catch block and I could see the exception was being caught the same way if it's placed within the inner using statement just as ShaneLS example.

     try
     {
       using (var con = new SqlConnection(@"Data Source=..."))
       {
         var cad = "INSERT INTO table VALUES (@r1,@r2,@r3)";

         using (var insertCommand = new SqlCommand(cad, con))
         {
           insertCommand.Parameters.AddWithValue("@r1", atxt);
           insertCommand.Parameters.AddWithValue("@r2", btxt);
           insertCommand.Parameters.AddWithValue("@r3", ctxt);
           con.Open();
           insertCommand.ExecuteNonQuery();
         }
       }
     }
     catch (Exception ex)
     {
       MessageBox.Show("Error: " + ex.Message, "UsingTest", MessageBoxButtons.OK, MessageBoxIcon.Error);
     }

No matter where's the try/catch placed, the exception will be caught without issues.


Yes to both questions. The using statement gets compiled into a try/finally block

using (SqlConnection connection = new SqlConnection(connectionString))
{
}

is the same as

SqlConnection connection = null;
try
{
    connection = new SqlConnection(connectionString);
}
finally
{
   if(connection != null)
        ((IDisposable)connection).Dispose();
}

Edit: Fixing the cast to Disposable http://msdn.microsoft.com/en-us/library/yh598w02.aspx


Dispose simply gets called when you leave the scope of using. The intention of "using" is to give developers a guaranteed way to make sure that resources get disposed.

From MSDN:

A using statement can be exited either when the end of the using statement is reached or if an exception is thrown and control leaves the statement block before the end of the statement.


Using generates a try / finally around the object being allocated and calls Dispose() for you.

It saves you the hassle of manually creating the try / finally block and calling Dispose()