Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

C# 8 switch expression with multiple cases with same result

I got around to installing it, but I have not found a way to specify multiple, separate case labels for a single switch section with the new syntax.

However, you can create a new variable that captures the value and then use a condition to represent the cases that should have the same result:

var resultText = switchValue switch
{
    var x when
        x == 1 ||
        x == 2 ||
        x == 3 => "one to three",
    4 => "four",
    5 => "five",
    _ => "unknown",
};

This is actually more concise if you have many cases to test, because you can test a range of values in one line:

var resultText = switchValue switch
{
    var x when x > 0 && x < 4 => "one to three",
    4 => "four",
    5 => "five",
    _ => "unknown",
};

C# 9 supports the following:

var switchValue = 3;
var resultText = switchValue switch
{
    1 or 2 or 3 => "one, two, or three",
    4 => "four",
    5 => "five",
    _ => "unknown",
};

Alternatively:

var switchValue = 3;
var resultText = switchValue switch
{
    >= 1 and <= 3 => "one, two, or three",
    4 => "four",
    5 => "five",
    _ => "unknown",
};

Source


For older versions of C#, I use the following extension method:

public static bool In<T>(this T val, params T[] vals) => vals.Contains(val);

like this:

var switchValue = 3;
var resultText = switchValue switch
{
    var x when x.In(1, 2, 3) => "one, two, or three",
    4 => "four",
    5 => "five",
    _ => "unknown",
};

It's a little more concise than when x == 1 || x == 2 || x == 3 and has a more natural ordering than when new [] {1, 2, 3}.Contains(x).


Sadly, this appears to be a shortcoming in the switch-expression syntax, relative to the switch-statement syntax. As other posters have suggested, the rather clumsy var syntax is your only real option.

So you might have been hoping you could write:

switchValue switch {
    Type1 t1:
    Type2 t2:
    Type3 t3 => ResultA, // where the ResultX variables are placeholders for expressions.
    Type4 t4 => ResultB,
    Type5 t5 => ResultC
};

Instead you will need to write the rather awkward code below, with typename sprayed about:

switchValue switch {
    var x when x is Type1 || x is Type2 || x is Type 3 => ResultA,
    Type4 t4 => ResultB,
    Type5 t5 => ResultC
};

In such a simple example, you can probably live with this awkwardness. But more complicated example are much less liveable with. In fact my examples are actually a simplification of an example drawn from our own code base, where I was hoping to convert a switch-statement, with roughly six outcomes but over a dozen type-cases, into a switch-expression. And the result was clearly less readable than the switch-statement.

My view is that if the switch-expression needs shared outcomes and is more than a few lines long, then you are better off sticking to a switch-statement. Boo! It's more verbose but probably a kindness to your teammates.

ResultType tmp;
switch (switchValue) {
    case Type1 t1:
    case Type2 t2:
    case Type3 t3:
        tmp = ResultA;
        break;
    case Type4 t4:
        tmp = ResultB;
        break;
    case Type5 t5:
        tmp = ResultC;
        break;
};
return tmp;