C# accuracy when checking float in List<float> with Contains method

Question

I have a list of floats and want to check if it already contains a particular value with the List.Contains() method. I know that for float equality tests you often can't use == but something like myFloat - value < 0.001.

My question is, does the Contains method account for this or I do I need to use a method that accounts for float precision errors for testing if the float is in the list?

Jordonjorey

Did you try it? Any issues met? – Diandiana

the Contains only does the simpliest comparison. it doesn't handle that. – Papeete

#3875127 – Unrequited

how about myList.Any(f=>Math.Abs(f-value)<0.0001) – Lemur

You actually can use == perfectly well with floats, you just need to know what that means... – Longford

@CoryNelson I strongly recommend reading twistedoakstudios.com/blog/… before assuming == is safe with floats. Specifically the part titled "Wrong #3". – Congresswoman

@BryceWagner thanks, that'll be useful to someone here. But my comment stands. Float comparison has become the new "never use goto". – Longford

Answer 1

From the docs for List(T).Contains:

This method determines equality by using the default equality comparer, as defined by the object's implementation of the IEquatable<T>.Equals method for T (the type of values in the list).

So you will need to handle comparison with a threshold yourself. For example, you can use your own custom equality comparer. Something like this:

public class FloatThresholdComparer : IEqualityComparer<float>
{
    private readonly float _threshold;
    public FloatThresholdComparer(float threshold)
    {
        _threshold = threshold;
    }

    public bool Equals(float x, float y)
    {
        return Math.Abs(x-y) < _threshold;
    }

    public int GetHashCode(float f)
    {
        throw new NotImplementedException("Unable to generate a hash code for thresholds, do not use this for grouping");
    }
}

And use it:

var result = floatList.Contains(100f, new FloatThresholdComparer(0.01f))

Valero

Answer 2

It just uses the default equality comparison for objects contained in the list. It will be the equivalent of calling object.Equals() when performing the comparisons.

If you need a different equality implementation, you could use the linq Contains() overload that accepts an equality comparer. Then you'd just have to implement that comparison and pass it in. This should perform roughly the same but ultimately slower.

Unhorse

Answer 3

The other answers are correct, but if you want an alternative quick-and-dirty solution without writing a new equality comparer, you could use the List.Exists Method:

bool found = list.Exists(num => Math.Abs(num - valueToFind) < 0.001);

Edit: My original answer said the above was Linq, however the Exists method is part of the List class. The same concept using Linq is below, using IEnumerable.Any:

bool found = list.Any(num => Math.Abs(num - valueToFind) < 0.001);

Malissamalissia

I like that - that's a clever application of LINQ – Millihenry