C# - how to determine whether a Type is a number

Question

Is there a way to determine whether or not a given .Net Type is a number? For example: System.UInt32/UInt16/Double are all numbers. I want to avoid a long switc

Answer 1

They are all value types (except for bool and maybe enum). So you could simply use:

bool IsNumberic(object o)
{
    return (o is System.ValueType && !(o is System.Boolean) && !(o is System.Enum))
}

Answer 2

Basically Skeet's solution but you can reuse it with Nullable types as follows:

public static class TypeHelper
{
    private static readonly HashSet<Type> NumericTypes = new HashSet<Type>
    {
        typeof(int),  typeof(double),  typeof(decimal),
        typeof(long), typeof(short),   typeof(sbyte),
        typeof(byte), typeof(ulong),   typeof(ushort),  
        typeof(uint), typeof(float)
    };
    
    public static bool IsNumeric(Type myType)
    {
       return NumericTypes.Contains(Nullable.GetUnderlyingType(myType) ?? myType);
    }
}

Answer 3

Type extension with null-type support.

public static bool IsNumeric(this Type type)
    {
        if (type == null) { return false; }

        if (type.IsGenericType && type.GetGenericTypeDefinition() == typeof(Nullable<>))
        {
            type = type.GetGenericArguments()[0];
        }

        switch (Type.GetTypeCode(type))
        {
            case TypeCode.Byte:
            case TypeCode.SByte:
            case TypeCode.UInt16:
            case TypeCode.UInt32:
            case TypeCode.UInt64:
            case TypeCode.Int16:
            case TypeCode.Int32:
            case TypeCode.Int64:
            case TypeCode.Decimal:
            case TypeCode.Double:
            case TypeCode.Single:
                return true;
            default:
                return false;
        }
    }

Answer 4

Switch is a little slow, bacause every time methods in the worst situation will be go through all type. I think, using Dictonary is more nice, in this situation you will be have O(1):

public static class TypeExtensions
{
    private static readonly HashSet<Type> NumberTypes = new HashSet<Type>();

    static TypeExtensions()
    {
        NumberTypes.Add(typeof(byte));
        NumberTypes.Add(typeof(decimal));
        NumberTypes.Add(typeof(double));
        NumberTypes.Add(typeof(float));
        NumberTypes.Add(typeof(int));
        NumberTypes.Add(typeof(long));
        NumberTypes.Add(typeof(sbyte));
        NumberTypes.Add(typeof(short));
        NumberTypes.Add(typeof(uint));
        NumberTypes.Add(typeof(ulong));
        NumberTypes.Add(typeof(ushort));
    }

    public static bool IsNumber(this Type type)
    {
        return NumberTypes.Contains(type);
    }
}

Answer 5

Approach based on Philip's proposal, enhanced with SFun28's inner type check for Nullable types:

public static class IsNumericType
{
    public static bool IsNumeric(this Type type)
    {
        switch (Type.GetTypeCode(type))
        {
            case TypeCode.Byte:
            case TypeCode.SByte:
            case TypeCode.UInt16:
            case TypeCode.UInt32:
            case TypeCode.UInt64:
            case TypeCode.Int16:
            case TypeCode.Int32:
            case TypeCode.Int64:
            case TypeCode.Decimal:
            case TypeCode.Double:
            case TypeCode.Single:
                return true;
            case TypeCode.Object:
                if (type.IsGenericType && type.GetGenericTypeDefinition() == typeof(Nullable<>))
                {
                    return Nullable.GetUnderlyingType(type).IsNumeric();
                    //return IsNumeric(Nullable.GetUnderlyingType(type));
                }
                return false;
            default:
                return false;
        }
    }
}

Why this? I had to check if a given Type type is a numeric type, and not if an arbitrary object o is numeric.

Answer 6

Don't use a switch - just use a set:

HashSet<Type> NumericTypes = new HashSet<Type>
{
    typeof(decimal), typeof(byte), typeof(sbyte),
    typeof(short), typeof(ushort), ...
};

EDIT: One advantage of this over using a type code is that when new numeric types are introduced into .NET (e.g. BigInteger and Complex) it's easy to adjust - whereas those types won't get a type code.