Determine if DataColumn is numeric

Question

Is there a better way than this to check if a DataColumn in a DataTable is numeric (coming from a SQL Server database)?

  Database db = DatabaseFactory.CreateDatabase();
  DbCommand cmd = db.GetStoredProcCommand("Get_Some_Data");
  DataSet ds = db.ExecuteDataSet(cmd);

  foreach (DataTable tbl in ds.Tables) {
    foreach (DataColumn col in tbl.Columns) {
      if (col.DataType == typeof(System.Single)
        || col.DataType == typeof(System.Double)
        || col.DataType == typeof(System.Decimal)
        || col.DataType == typeof(System.Byte)
        || col.DataType == typeof(System.Int16)
        || col.DataType == typeof(System.Int32)
        || col.DataType == typeof(System.Int64)) {
        // this column is numeric
      } else {
        // this column is not numeric
      }
    }
  }

Answer 1

There is no good way to check if the type is numeric except comparing it to the actual types.
This is especially true if the definition of numeric is a bit different (in your case, according to code, - unsigned integers are not numerics).

Another thing is that DataColumn.DataType according to MSDN only supports following types:

• Boolean
• Byte
• Char
• DateTime
• Decimal
• Double
• Int16
• Int32
• Int64
• SByte
• Single
• String
• TimeSpan
• UInt16
• UInt32
• UInt64
• Byte[]

The bolded types are numerics (as I define it) so you need to make sure you check them.

I personally would write an extension method for the DataColumn type (not for the TYPE!).
I hate the if...then..else thing so instead I use a SETS-based approach, like this:

public static bool IsNumeric(this DataColumn col) {
  if (col == null)
    return false;
  // Make this const
  var numericTypes = new [] { typeof(Byte), typeof(Decimal), typeof(Double),
        typeof(Int16), typeof(Int32), typeof(Int64), typeof(SByte),
        typeof(Single), typeof(UInt16), typeof(UInt32), typeof(UInt64)};
  return numericTypes.Contains(col.DataType);
}

And the usage would be:

if (col.IsNumeric()) ....

which is easy enough for me