How can I compare a string to a "filter" list in linq?

Question

I'm trying to filter a collection of strings by a "filter" list... a list of bad words. The string contains a word from the list I dont want it.

I've gotten so far, the bad Word here is "frakk":

string[] filter = { "bad", "words", "frakk" };

string[] foo = 
{ 
    "this is a lol string that is allowed", 
    "this is another lol frakk string that is not allowed!"
};

var items = from item in foo 
            where (item.IndexOf( (from f in filter select f).ToString() ) == 0)
            select item;

But this aint working, why?

Answer 1

You can use Any + Contains:

var items = foo.Where(s => !filter.Any(w => s.Contains(w)));

if you want to compare case-insensitively:

var items = foo.Where(s => !filter.Any(w => s.IndexOf(w, StringComparison.OrdinalIgnoreCase) >= 0));

Update: If you want to exlude sentences where at least one word is in the filter-list you can use String.Split() and Enumerable.Intersect:

var items = foo.Where(sentence => !sentence.Split().Intersect(filter).Any());

Enumerable.Intersect is very efficient since it uses a Set under the hood. it's more efficient to put the long sequence first. Due to Linq's deferred execution is stops on the first matching word.

( note that the "empty" Split includes other white-space characters like tab or newline )

Answer 2

The first problem you need to solve is breaking up the sentence into a series of words. The simplest way to do this is based on spaces

string[] words = sentence.Split(new char[] {' '}, StringSplitOptions.RemoveEmptyEntries);

From there you can use a simple LINQ expression to find the profanities

var badWords = words.Where(x => filter.Contains(x));

However this is a bit of a primitive solution. It won't handle a number of complex cases that you likely need to think about

• There are many characters which qualify as a space. My solution only uses ' '
• The split doesn't handle punctuations. So dog! won't be viewed as dog. Probably much better to break up words on legal characters