2

Is there a fast/simple way to calculate the frequency distribution of a .Net collection using Linq or otherwise?

For example: An arbitrarily long List contains many repetitions. What's a clever way of walking the list and counting/tracking repetitions?

Paul Sasik
  • 79,492
  • 20
  • 149
  • 189

3 Answers3

5

The simplest way to find duplicate items in a list is to group it, like this:

var dups = list.GroupBy(i => i).Where(g => g.Skip(1).Any());

(Writing Skip(1).Any() should be faster than (Count() > 1) because it won't have to traverse more than two items from each group. However, the difference is probably negligible unless list's enumerator is slow)

SLaks
  • 868,454
  • 176
  • 1,908
  • 1,964
2

The easiest way is to use a hashmap and either use the value as the key and increment the value, or pick a bucket size (bucket 1 = 1 - 10, bucket 2 = 11 - 20, etc), and increment each bucket by the value.

Then you can go through and determine the frequencies.

James Black
  • 41,583
  • 10
  • 86
  • 166
1

The C5 generic collections library has a HashBag implementation that accepts duplicates by counting. The following pseudo-code would get you what you're looking for:

var hash = new HashBag();
hash.AddAll(list);
var mults = hash.ItemMultiplicities();

(where K is the type of the items in your list) mults will then contain an IDictionary<K,int> where the list item is the key and the multiplicity is the value.

Marcus Griep
  • 8,156
  • 1
  • 23
  • 24