I'm looking for an efficient way of sorting an array of email addresses to avoid items with the same domain to be consecutive, in C#.
Email addresses inside the array are already distinct and all of them are lower case.
Example:
Given an array with the following entries:
[email protected]
[email protected]
[email protected]
[email protected]
I would like to obtain something similar to the following:
[email protected]
[email protected]
[email protected]
[email protected]
With the help of an extension method (stolen from https://stackoverflow.com/a/27533369/172769), you can go like this:
List<string> emails = new List<string>();
emails.Add("[email protected]");
emails.Add("[email protected]");
emails.Add("[email protected]");
emails.Add("[email protected]");
var q = emails.GroupBy(m => m.Split('@')[1]).Select(g => new List<string>(g)).Interleave();
The Interleave method is defined as:
public static IEnumerable<T> Interleave<T>(this IEnumerable<IEnumerable<T>> source )
{
var queues = source.Select(x => new Queue<T>(x)).ToList();
while (queues.Any(x => x.Any())) {
foreach (var queue in queues.Where(x => x.Any())) {
yield return queue.Dequeue();
}
}
}
So basically, we create groups based on the domain part of the email adresses, project (or Select) each group into a List<string>, and then "Interleave" those lists.
I have tested against your sample data, but more thorough testing might be needed to find edge cases.
Cheers
Something like this will spread them equally, but you will have the problems (=consecutive elements) at the end of the new list...
var list = new List<string>();
list.Add("[email protected]");
list.Add("[email protected]");
list.Add("[email protected]");
list.Add("[email protected]");
var x = list.GroupBy(content => content.Split('@')[1]);
var newlist = new List<string>();
bool addedSomething=true;
int i = 0;
while (addedSomething) {
addedSomething = false;
foreach (var grp in x) {
if (grp.Count() > i) {
newlist.Add(grp.ElementAt(i));
addedSomething = true;
}
}
i++;
}
This will distribute them semi-evenly and attempt to avoid matching domains next to each other (although in certain lists that may be impossible). This answer will use OOP and Linq.
using System;
using System.Linq;
using System.Collections.Generic;
public class Program
{
public static void Main()
{
var seed = new List<string>()
{
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]",
"[email protected]"
};
var work = seed
// Create a list of EmailAddress objects
.Select(s => new EmailAddress(s)) // s.ToLowerCase() ?
// Group the list by Domain
.GroupBy(s => s.Domain)
// Create a List<EmailAddressGroup>
.Select(g => new EmailAddressGroup(g))
.ToList();
var currentDomain = string.Empty;
while(work.Count > 0)
{
// this list should not be the same domain we just used
var noDups = work.Where(w => w.Domain != currentDomain);
// if none exist we are done, or it can't be solved
if (noDups.Count() == 0)
{
break;
}
// find the first group with the most items
var workGroup = noDups.First(w => w.Count() == noDups.Max(g => g.Count()));
// get the email address and remove it from the group list
var workItem = workGroup.Remove();
// if the group is empty remove it from *work*
if (workGroup.Count() == 0)
{
work.Remove(workGroup);
Console.WriteLine("removed: " + workGroup.Domain);
}
Console.WriteLine(workItem.FullEmail);
// last domain looked at.
currentDomain = workItem.Domain;
}
Console.WriteLine("Cannot disperse email addresses affectively, left overs:");
foreach(var workGroup in work)
{
while(workGroup.Count() > 0)
{
var item = workGroup.Remove();
Console.WriteLine(item.FullEmail);
}
}
}
public class EmailAddress
{
public EmailAddress(string emailAddress)
{
// Additional Email Address Validation
var result = emailAddress.Split(new char[] {'@'}, StringSplitOptions.RemoveEmptyEntries)
.ToList();
if (result.Count() != 2)
{
new ArgumentException("emailAddress");
}
this.FullEmail = emailAddress;
this.Name = result[0];
this.Domain = result[1];
}
public string Name { get; private set; }
public string Domain { get; private set; }
public string FullEmail { get; private set; }
}
public class EmailAddressGroup
{
private List<EmailAddress> _emails;
public EmailAddressGroup(IEnumerable<EmailAddress> emails)
{
this._emails = emails.ToList();
this.Domain = emails.First().Domain;
}
public int Count()
{
return _emails.Count();
}
public string Domain { get; private set; }
public EmailAddress Remove()
{
var result = _emails.First();
_emails.Remove(result);
return result;
}
}
}
Output:
removed: b.com
removed: c.com
Cannot disperse email addresses affectively, left overs:
Edit: Added a high level description :)
What this code does is group each element by the domain, sort the groups by size in descending order (largest group first), project the elements of each group into a stack, and pop them off of each stack (always pop the next element off the largest stack with a different domain). If there is only a single stack left, then its contents are yielded.
This should make sure that all domains distributed as evenly as possible.
MaxBy extension method from: https://stackoverflow.com/a/31560586/969962
private IEnumerable<string> GetNonConsecutiveEmails(List<string> list)
{
var emailAddresses = list.Distinct().Select(email => new EmailAddress { Email = email, Domain = email.Split('@')[1]}).ToArray();
var groups = emailAddresses
.GroupBy(addr => addr.Domain)
.Select (group => new { Domain = group.Key, EmailAddresses = new Stack<EmailAddress>(group)})
.ToList();
EmailAddress lastEmail = null;
while(groups.Any(g => g.EmailAddresses.Any()))
{
// Try and pick from the largest stack.
var stack = groups
.Where(g => (g.EmailAddresses.Any()) && (lastEmail == null ? true : lastEmail.Domain != g.Domain))
.MaxBy(g => g.EmailAddresses.Count);
// Null check to account for only 1 stack being left.
// If so, pop the elements off the remaining stack.
lastEmail = (stack ?? groups.First(g => g.EmailAddresses.Any())).EmailAddresses.Pop();
yield return lastEmail.Email;
}
}
class EmailAddress
{
public string Domain;
public string Email;
}
public static class Extensions
{
public static T MaxBy<T,U>(this IEnumerable<T> data, Func<T,U> f) where U:IComparable
{
return data.Aggregate((i1, i2) => f(i1).CompareTo(f(i2))>0 ? i1 : i2);
}
}
What I am trying to do here is to sort them first. Then I re-arrange from a different end. I'm sure there're more efficient ways to do this but this is one easy way to do it.
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
namespace ConsoleApplication4
{
class Program
{
static void Main(string[] args)
{
String[] emails = { "[email protected]", "[email protected]", "[email protected]", "[email protected]" };
var result = process(emails);
}
static String[] process(String[] emails)
{
String[] result = new String[emails.Length];
var comparer = new DomainComparer();
Array.Sort(emails, comparer);
for (int i = 0, j = emails.Length - 1, k = 0; i < j; i++, j--, k += 2)
{
if (i == j)
result[k] = emails[i];
else
{
result[k] = emails[i];
result[k + 1] = emails[j];
}
}
return result;
}
}
public class DomainComparer : IComparer<string>
{
public int Compare(string left, string right)
{
int at_pos = left.IndexOf('@');
var left_domain = left.Substring(at_pos, left.Length - at_pos);
at_pos = right.IndexOf('@');
var right_domain = right.Substring(at_pos, right.Length - at_pos);
return String.Compare(left_domain, right_domain);
}
}
}
