Question

I’m writing database test utils code which compares DataRow values with expected values supplied as a Dictionary (columnName, expectedColumnValue).

It works fine for many types, but for byte and short I had to add conversion code that converts the values to an Int32.

Two questions:

1. Do you have ideas how to make this code better, i.e. how to avoid the conversion?
   For decimal and float it seems to work as they are explicitly declared as a decimal/float value. Long also works without problems.

2. If there is no other way, are there any other types apart from short and byte that I need to worry about?

Demo code below:

            var table = new DataTable();
            table.Columns.Add(new DataColumn("CarrierId", typeof(byte)));
            table.Columns.Add(new DataColumn("NotationId", typeof(short)));

            var row = table.NewRow();
            row[0] = 5;
            row[1] = 123;

            table.Rows.Add(row);

            var expected = new Dictionary<string, object>
            {
                {"CarrierId", 5},
                {"NotationId", 123},
            };

            foreach (var entry in expected)
            {
                var value = row[entry.Key];
                var expectedValue = entry.Value;

                if (value is short || value is byte)
                    value = Convert.ToInt32(value);

                Console.WriteLine();
                Console.WriteLine(String.Format("Not converted: {0}", row[entry.Key].Equals(entry.Value)));
                Console.WriteLine(String.Format("Converted (if applicable): {0}", value.Equals(expectedValue)));

            }

Answer 1

Add the values in their expected type to the dictionary, then you should not need to do the conversion any more when comparing:

var expected = new Dictionary<string, object>
{
    { "CarrierId", (byte)5 },
    { "NotationId", (short)123 },
};

---

When you assign an int number given by a constant expression (known at compile time) to a byte variable, C# automatically converts it to byte.

byte b = 5; // Stores a byte

But when you assign it to an object variable, C# does not know that you will be using it as a byte in future and threats it as an int by default.

object o = 5; // Stores an int