How to traverse keys of a Hashtable in alphabetical order?

Question

What is the easiest way to traverse a hashtable's keys in ascending alphabetical order?

Answer 1

This is fairly dependent upon what the type of the key is. But lets assume for a minute that they are strings. You could use the following LINQ query

Hashtable table = GetHashTable();
var keys = table.Keys.Cast<String>().OrderBy(x => x);

For more complex structures the LINQ query is only slightly different. Lets assume you had the following definition for a key

struct Name {
  public string First;
  public string Last;
  // Equality code omitted
}

The LINQ code would be the following

Hashtable table = GetHashtable();
var keys = table.Keys.Cast<Name>().OrderBy(x => x.First).ThenBy(x => x.Last);

Answer 2

Thats not really what hash tables are designed for (they are made to have uniform distribution of keys). Use a sorted tree?

Answer 3

If you want a map which keeps its keys in natural order, I suggest you don't use Hashtable to start with. If you're still using 1.1, using System.Collections.SortedList. If you're using 2.0 or higher, use SortedList<TKey, TValue> or SortedDictionary<TKey, TValue>. The latter two are largely the same in terms of API, but have different performance characteristics - see the docs for more information.

Answer 4

Well, I found this snippet to be the most suitable to my situation:

Hashtable settings = GetSettings();
ArrayList keys = new ArrayList();
keys.AddRange(settings.Keys);
keys.Sort();
foreach (object key in keys)
{
    // Logic here
}

Answer 5

It'll probably be slightly faster to use SortedList -

SortedList settings = new SortedList(GetSettings());
foreach (object key in settings.Keys)
{
    //logic
}

creating & sorting the ArrayList is O(n) + O(nlog n) = O(nlog n), whereas the SortedList constructor (according to the docs) is O(n), so it'll be faster to use SortedList directly rather than using an arraylist and explicitly sorting