C# List-Sampling Extension Method

While messing around with some for-fun code in C# today, I found myself looking for something along the lines of Ruby's Array#sample method, but for .NET's IList. Strangely (or perhaps not-so-strangely, because when you search for something like "C# IList sample extension", you get tons of noise), I couldn't find anything. So I wrote a simple one to share here:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

public static class ListExtensions { readonly static Random myRand = new Random(); readonly static object lockObj = new object(); public static T Sample<T>(this IList<T> list) { lock (lockObj) { return Sample(list, myRand); } } public static T Sample<T>(this IList<T> list, Random rand) { if (list == null || list.Count == 0) { return default(T); } return list[rand.Next(list.Count)]; } public static IList<T> Sample<T>(this IList<T> list, int size) { lock (lockObj) { return Sample(list, size, myRand); } } public static IList<T> Sample<T>(this IList<T> list, int size, Random rand) { if (list == null) { return new List<T>(); } return list.Select(x => new { Entropy = rand.Next(), Item = x }) .OrderBy(x => x.Entropy) .Select(x => x.Item) .Take(size) .ToList(); } }

Notes:

• I put this on the IList interface instead of, say, IEnumerable, because the random access helps greatly with the efficiency of the size-1 overloads. (And that just happens to be the specific case I needed today).
• You can use the overloads that require you to pass your own Random instance, or not, but if you choose not to, note that you'll take a locking penalty since System.Random isn't thread-safe.
• Samples of size > 1 are implemented via that Linq-style shuffle.