What is a better alternative to comparing string.tolower?

Question

from my past experience and stackoverflow, I learned that String.ToLower() is bad in performance. Now I have the following issue, I need to filter out or do a specific action when specific words are in a giant list.

Old approach, which I want to fix:

if (input.Any(i => i.ToLower() == "alle" || i.ToLower() == "all" || i.ToLower() == "none") 
{
    // do something
}

I was thinking of using a hashset, but I am questioning the performance and how it handles the case sensitivity, I basically dont care about the case sensitivity. Does it make sense for me to use the hashset?

my current suggestion as a solution:

var unwantedInputsSet = new HashSet<string> {"alle", "all", "none"};
if (input.Any(i => i => unwantedInputsSet.Contains(i))) 
{
    // do something
}

Is there any better alternative to this or not. Do you have any ideas how to approach this better?

Answer 1

You can pass comparer to the HashSet, for example StringComparer.InvariantCultureIgnoreCase:

var unwantedInputsSet = new HashSet<string>(StringComparer.InvariantCultureIgnoreCase) {"alle", "all", "none"};
if (input.Any(i => unwantedInputsSet.Contains(i))) 
{
    // do something
}

Or, as suggested in comments, use pattern matching:

if (input.Any(i => i.ToLower() is "alle" or "all" or "none") 
{
    // do something
}

Which should be turned by compiler into code similar to yours (though ToLower should be called once).

As for performance - it can be highly depended on the actual data and you should measure it using expected datasets. For small search set HashSet can be performing worse than several comparisons like:

var cmpr = StringComparison.InvariantCultureIgnoreCase;
if (input.Any(i => string.Equals(i, "alle", cmpr) || string.Equals(i, "all", cmpr) || string.Equals(i, "none", cmpr)))
{
    // do something
}

For such benchmarking I recommend looking into BenchmarkDotNet.