We Keep Coding

algorithm for selecting N random elements from a List<T> in C# [duplicate]

This question already has answers here:
Randomize a List<T>
(28 answers)
Closed 6 years ago.
I need a quick algorithm to select 4 random elements from a generic list. For example, I'd like to get 4 random elements from a List and then based on some calculations if elements found not valid then it should again select next 4 random elements from the list.

You could do it like this
public static class Extensions
{
public static Dictionary<int, T> GetRandomElements<T>(this IList<T> list, int quantity)
{
var result = new Dictionary<int, T>();
if (list == null)
return result;
Random rnd = new Random(DateTime.Now.Millisecond);
for (int i = 0; i < quantity; i++)
{
int idx = rnd.Next(0, list.Count);
result.Add(idx, list[idx]);
}
return result;
}
}
Then use the extension method like this:
List<string> list = new List<string>() { "a", "b", "c", "d", "e", "f", "g", "h" };
Dictionary<int, string> randomElements = list.GetRandomElements(3);
foreach (KeyValuePair<int, string> elem in randomElements)
{
Console.WriteLine($"index in original list: {elem.Key} value: {elem.Value}");
}

something like that:
using System;
using System.Collections.Generic;
public class Program
{
public static void Main()
{
var list = new List<int>();
list.Add(1);
list.Add(2);
list.Add(3);
list.Add(4);
list.Add(5);
int n = 4;
var rand = new Random();
var randomObjects = new List<int>();
for (int i = 0; i<n; i++)
{
var index = rand.Next(list.Count);
randomObjects.Add(list[index]);
}
}
}

You can store indexes in some list to get non-repeated indexes:
List<T> GetRandomElements<T>(List<T> allElements, int randomCount = 4)
{
if (allElements.Count < randomCount)
{
return allElements;
}
List<int> indexes = new List<int>();
// use HashSet if performance is very critical and you need a lot of indexes
//HashSet<int> indexes = new HashSet<int>();
List<T> elements = new List<T>();
Random random = new Random();
while (indexes.Count < randomCount)
{
int index = random.Next(allElements.Count);
if (!indexes.Contains(index))
{
indexes.Add(index);
elements.Add(allElements[index]);
}
}
return elements;
}
Then you can do some calculation and call this method:
void Main(String[] args)
{
do
{
List<int> elements = GetRandomelements(yourElements);
//do some calculations
} while (some condition); // while result is not right
}

Suppose that the length of the List is N. Now suppose that you will put these 4 numbers in another List called out. Then you can loop through the List and the probability of the element you are on being chosen is
(4 - (out.Count)) / (N - currentIndex)

funcion (list)
(
loop i=0 i < 4
index = (int) length(list)*random(0 -> 1)
element[i] = list[index]
return element
)
while(check == false)
(
elements = funcion (list)
Do some calculation which returns check == false /true
)
This is the pseudo code, but i think you should of come up with this yourself.
Hope it helps:)

All the answers up to now have one fundamental flaw; you are asking for an algorithm that will generate a random combination of n elements and this combination, following some logic rules, will be valid or not. If its not, a new combination should be produced. Obviously, this new combination should be one that has never been produced before. All the proposed algorithms do not enforce this. If for example out of 1000000 possible combinations, only one is valid, you might waste a whole lot of resources until that particular unique combination is produced.
So, how to solve this? Well, the answer is simple, create all possible unique solutions, and then simply produce them in a random order. Caveat: I will suppose that the input stream has no repeating elements, if it does, then some combinations will not be unique.
First of all, lets write ourselves a handy immutable stack:
class ImmutableStack<T> : IEnumerable<T>
{
public static readonly ImmutableStack<T> Empty = new ImmutableStack<T>();
private readonly T head;
private readonly ImmutableStack<T> tail;
public int Count { get; }
private ImmutableStack()
{
Count = 0;
}
private ImmutableStack(T head, ImmutableStack<T> tail)
{
this.head = head;
this.tail = tail;
Count = tail.Count + 1;
}
public T Peek()
{
if (this == Empty)
throw new InvalidOperationException("Can not peek a empty stack.");
return head;
}
public ImmutableStack<T> Pop()
{
if (this == Empty)
throw new InvalidOperationException("Can not pop a empty stack.");
return tail;
}
public ImmutableStack<T> Push(T item) => new ImmutableStack<T>(item, this);
public IEnumerator<T> GetEnumerator()
{
var current = this;
while (current != Empty)
{
yield return current.head;
current = current.tail;
}
}
IEnumerator IEnumerable.GetEnumerator() => GetEnumerator();
}
This will make our life easier while producing all combinations by recursion. Next, let's get the signature of our main method right:
public static IEnumerable<IEnumerable<T>> GetAllPossibleCombinationsInRandomOrder<T>(
IEnumerable<T> data, int combinationLength)
Ok, that looks about right. Now let's implement this thing:
var allCombinations = GetAllPossibleCombinations(data, combinationLength).ToArray();
var rnd = new Random();
var producedIndexes = new HashSet<int>();
while (producedIndexes.Count < allCombinations.Length)
{
while (true)
{
var index = rnd.Next(allCombinations.Length);
if (!producedIndexes.Contains(index))
{
producedIndexes.Add(index);
yield return allCombinations[index];
break;
}
}
}
Ok, all we are doing here is producing random indexees, checking we haven't produced it yet (we use a HashSet<int> for this), and returning the combination at that index.
Simple, now we only need to take care of GetAllPossibleCombinations(data, combinationLength).
Thats easy, we'll use recursion. Our bail out condition is when our current combination is the specified length. Another caveat: I'm omitting argument validation throughout the whole code, things like checking for null or if the specified length is not bigger than the input length, etc. should be taken care of.
Just for the fun, I'll be using some minor C#7 syntax here: nested functions.
public static IEnumerable<IEnumerable<T>> GetAllPossibleCombinations<T>(
IEnumerable<T> stream, int length)
{
return getAllCombinations(stream, ImmutableStack<T>.Empty);
IEnumerable<IEnumerable<T>> getAllCombinations<T>(IEnumerable<T> currentData, ImmutableStack<T> combination)
{
if (combination.Count == length)
yield return combination;
foreach (var d in currentData)
{
var newCombination = combination.Push(d);
foreach (var c in getAllCombinations(currentData.Except(new[] { d }), newCombination))
{
yield return c;
}
}
}
}
And there we go, now we can use this:
var data = "abc";
var random = GetAllPossibleCombinationsInRandomOrder(data, 2);
foreach (var r in random)
{
Console.WriteLine(string.Join("", r));
}
And sure enough, the output is:
bc
cb
ab
ac
ba
ca

C# How to work-around identical keys in a Dictionary

This program add's values with their other values into a dictionary, all is fine until there are identical keys (var eat(.8) and var extra(.8) with different values. How do i ensure that i can use the right key every time even though they are similar? For example, var example = gigantDictionary[.8] (but i want var edmg value instead of '500' in the code?
var wqat = 1.1; //| index 0
var rat = .2; //| index 1
var eat = .8; //| index 2
var baat = 1.2; //| index 3
var extra = .8; //| index 4
var wqdmg = 120; //| index 0
var rdmg = 60; //| index 1
var edmg = 50; //| index 2
var badmg = 40; //| index 3
var extradmg = 500; //| index 4
List<double> theOneList = new List<double>();
List<double> damageList = new List<double>();
theOneList.Add(wqat);
theOneList.Add(rat);
theOneList.Add(eat);
theOneList.Add(baat);
damageList.Add(wqdmg);
damageList.Add(edmg);
damageList.Add(rdmg);
damageList.Add(badmg);
Dictionary<double, double> gigantDictionary = new Dictionary<double, double>();
for (int i = 0; i < theOneList.Count; i++)
{
gigantDictionary.Add(theOneList[i], damageList[i]);
gigantDictionary.Add(extra, 500); //this is the malignant similar key
}
theOneList.Sort((c, p) => -c.CompareTo(p)); //orders the list
List<double> finalList = new List<double>();
for (int i = 0; i < theOneList.Count; i++)
{
finalList.Add(gigantDictionary[theOneList[i]]); //grabs damage values and add's it to 'finalList'
Console.WriteLine(finalList[i]);
}
So ultimately, i want to order 'theOneList' by descent, in doing so i can get the damages from 'gigantDictionary' and put those into 'finalList', now i have an ordered damage list that i need, but since 2 keys are similar... this is holding me back.
*Edit: Could identical indexes be the key to this? be the bridge? for example, in index 0 i get 1.1 and 120, maybe the answer lies with the identical indexes, i want to get '120' damage from '1.1', notice both have index 0, this might work

They keys aren't "similar" they're "identical". If the keys were just "similar" then, as far as the dictionary is concerned, it's no different than being "completely different". From the point of view of a dictionary items are either equal, or not equal.
For example,
var example = gigantDictionary[.8]
(but i want var edmg value instead of '500' in the code?)
But how should the dictionary possibly know that? How would it know if you actually wanted to get 500?
Do you want to prevent the duplicate keys from being added, and instead always use the first value paired with every key? If so, just check if a key exists before adding a new one.
Do you want to just get all values associated with a key, if there are duplicates? Then have a dictionary where the value is a collection, and add all values associated with that one key to the collection.
Is there actually some way to distinguish the keys so that they're not actually identical? Then do that. With just a double (which is a very bad type to use as a key for a dictionary in the first place, as it's easy for floating point rounding errors to result in similar but different doubles that you consider equivalent) there's no good way to do this, but if your actual key could be different in such a way that distinguishes the two keys, then each could point to a unique value.

Right now you have two separate list for values that must go together. A better approach is to create a structure with the two values and keep a single list.
public class Thing
{
public string Name { get; set; }
public double TheOne { get; set; }
public double Dmg { get; set; }
}
class Program
{
static void Main(string[] args)
{
List<Thing> list=new List<Thing>() {
new Thing() { Name = "wq", TheOne = 1.1, Dmg=120 },
new Thing() { Name = "r", TheOne = 0.2, Dmg=60 },
new Thing() { Name = "e", TheOne = 0.8, Dmg=50 },
new Thing() { Name = "ba", TheOne = 1.2, Dmg=40 },
new Thing() { Name = "extra", TheOne = 0.8, Dmg=500 },
};
list.Sort((t1, t2) => t1.TheOne.CompareTo(t2.TheOne));
double[] dmg_list=list.Select((t) => t.Dmg).ToArray();
}
}
Edit 1
A constructor to Thing can be used to assign values with one operation.
public class Thing
{
// Constructor sets all the values
public Thing(string name, double theone, double dmg)
{
this.Name=name;
this.TheOne=theone;
this.Dmg=dmg;
}
public string Name { get; private set; }
public double TheOne { get; private set; }
public double Dmg { get; private set; }
}
class Program
{
static void Main(string[] args)
{
List<Thing> list=new List<Thing>();
list.Add(new Thing("wq", 1.1, 120));
list.Add(new Thing("r", 0.2, 60));
list.Add(new Thing("e", 0.8, 50));
list.Add(new Thing("ba", 1.2, 40));
list.Add(new Thing("extra", 0.8, 500));
list.Sort((t1, t2) => t1.TheOne.CompareTo(t2.TheOne));
double[] dmg_list=list.Select((t) => t.Dmg).ToArray();
}
}

Adding a value to a specific location in a list of queues

Queues:
public class Queue
{
public Queue() { }
public process Front() { return this.q_list.ElementAt(0); }
public int Size { get { return this.q_list.Count; } }
public bool IsEmpty { get { return this.q_list.Count <= 0 ? true : false; } }
public void Enqueue(process proc) { this.q_list.Add(proc); }
public void Dequeue() { this.q_list.RemoveAt(0); }
public List<process> q_list = new List<process>();
};
Creation of a list:
List<Queue> rr_list = new List<Queue>();
The process struct:
public class process
{
public int Proc_a;
public int Proc_b;
public int Proc_Index;
};
Let's say I want to add a process to the list at a specific location depending on the value of Proc_Index. How can I do that? Let's also assume the list is initially empty.
process proc = new process{
Proc_a = 1,
Proc_b = 2,
Proc_Index = 4 };
I want to add that to a queue that is in the list located at index 4.
Is this possible?
I've tried:
rr_list[proc.Proc_Index].Enqueue(proc);
But it says there's an issue with index not being found or something.
The only thing I can thing of is initializing the list by adding empty queues for up to 20 indexes, but I don't know if there's a better way.

You should use a System.Collections.Generic.Queue instead of writing your own. Use a System.Collections.Generic.Dictionary if you want key-value lookup.
var rr_list = new Dictionary<int, Queue<process>>();
process proc = new process{
Proc_a = 1,
Proc_b = 2,
Proc_Index = 4 };
rr_list[proc.Proc_Index].Enqueue(proc);

You may want to use a dictionary instead of a list.
var rr_list = new Dictionary<int, Queue>();
Then have an addprocess function as such
function void AddProcess(proccess proc){
if(rr_list.ContainsKey(proc.Proc_Index){
rr_list[proc.Proc_Index].Enqueue(proc);
} else {
rr_list[proc.Proc_Index] = (new Queue()).Enqueue(proc);
}
}

A list is usually supposed to have no holes, so if you were to add an element at index 4 to an empty list, this would make indexes 0 to 3 contain null.
Now, you can do it like that. You could check if the length is bigger than the requested index, and if not, keep adding null values until it is. Then the index would exist, and you could assign something to it:
static void EnsureLength<T> (List<T> list, int index)
{
while (list.Count <= index)
list.Add(default(T));
}
Then you could use it like this:
List<int?> list = new List<int?>();
EnsureLength(list, 3);
list[3] = 123;
A possibly better way would be to simply use a Dictionary, especially if you know that you will have holes. So you would just have a Dictionary<int, T>:
Dictionary<int, int?> dict = new Dictionary<int, int?>();
dict[3] = 123;

Creating a two-dimensional array

I am trying to create a two dimensional array and I am getting so confused. I was told by a coworker that I need to create a dictionary within a dictionary for the array list but he couldn't stick around to help me.
I have been able to create the first array that lists the the programs like this
+ project 1
+ project 2
+ project 3
+ project 4
The code that accomplishes this task is below-
var PGList = from x in db.month_mapping
where x.PG_SUB_PROGRAM == SP
select x;
//select x.PG.Distinct().ToArray();
var PGRow = PGList.Select(x => new { x.PG }).Distinct().ToArray();
So that takes care of my vertical array and now I need to add my horizontal array so that I can see the total amount spent in each accounting period. So the final output would look like this but without the dashes of course.
+ program 1-------100---200---300---400---500---600---700---800---900---1000---1100---1200
+ program 2-------100---200---300---400---500---600---700---800---900---1000---1100---1200
+ program 3-------100---200---300---400---500---600---700---800---900---1000---1100---1200
+ program 4-------100---200---300---400---500---600---700---800---900---1000---1100---1200
I have tried to use a foreach to cycle through the accounting periods but it doesn't work. I think I might be on the right track and I was hoping SO could provide some guidance or at the very least a tutorial for me to follow. I have posted the code that I written so far on the second array below. I am using C# and MVC 3. You might notice that their is no dictionary within a dictionary. If my coworker is correct how would I do something like that, I took a look at this question using dictionary as a key in other dictionary but I don't understand how I would use it in this situation.
Dictionary<string, double[]> MonthRow = new Dictionary<string, double[]>();
double[] PGContent = new double[12];
string lastPG = null;
foreach (var item in PGRow)
{
if (lastPG != item.PG)
{
PGContent = new double[12];
}
var MonthList = from x in db.Month_Web
where x.PG == PG
group x by new { x.ACCOUNTING_PERIOD, x.PG, x.Amount } into pggroup
select new { accounting_period = pggroup.Key.ACCOUNTING_PERIOD, amount = pggroup.Sum(x => x.Amount) };
foreach (var P in MonthList)
{
int accounting_period = int.Parse(P.accounting_period) - 1;
PAContent[accounting_period] = (double)P.amount;
MonthRow[item.PG] = PGContent;
lastPG = item.PG;
}
I hope I have clearly explained my issue, please feel free to ask for any clarification needed as I need to solve this problem and will be checking back often. Thanks for your help!

hope this helps.
// sample data
var data = new Dictionary<string, List<int>>();
data.Add("program-1", new List<int>() { 100, 110, 130 });
data.Add("program-2", new List<int>() { 200, 210, 230 });
data.Add("brogram-3", new List<int>() { 300, 310, 330 });
// query data
var newData = (from x in data
where x.Key.Contains("pro")
select x).ToDictionary(v => v.Key, v=>v.Value);
// display selected data
foreach (var kv in newData)
{
Console.Write(kv.Key);
foreach (var val in kv.Value)
{
Console.Write(" ");
Console.Write(val.ToString());
}
Console.WriteLine();
}
output is:
program-1 100 110 130
program-2 200 210 230

Don't try to use anonymous types or LINQ projection to create new data types, especially if you're a beginner, you will just get confused. If you want a specialized data type, define one; e.g.:
public class Account
{
public string Name { get; private set; }
public decimal[] MonthAmount { get; private set; }
readonly int maxMonths = 12;
public Account(string name, ICollection<decimal> monthAmounts)
{
if (name == null)
throw new ArgumentNullException("name");
if (monthAmounts == null)
throw new ArgumentNullException("monthAmounts");
if (monthAmounts.Count > maxMonths)
throw new ArgumentOutOfRangeException(string.Format(" monthAmounts must be <= {0}", maxMonths));
this.Name = name;
this.MonthAmount = new decimal[maxMonths];
int i = 0;
foreach (decimal d in monthAmounts)
{
this.MonthAmount[i] = d;
i++;
}
}
}
Use instances of this type directly, you do not have to convert them to arrays, dictionaries, lists, or anything else:
var accountPeriods = new List<Account>();
accountPeriods.Add(new Account("program-1", new decimal[] { 1, 2, 3, 4 }));
You can use LINQ or whatever to query or alter instances of your new type:
foreach (Account a in accountPeriods)
foreach (decimal d in a.MonthAmount)
DoSomethingWith(d);
That should be enough to get you started.

I want to thank #Ray Cheng and #Dour High Arch for their help but I have figured out another way to accomplish this task and I wanted to post my code so that the next person that is having the same trouble can figure out their problem faster.
Above I split my code into more managable sections to explain my problem as clearly as I could and the code below has all those parts combined so you can see the big picture. This code returns an array that contains the program and the amounts for every month.
public virtual ActionResult getAjaxPGs(string SP = null)
{
if (SP != null)
{
var PGList = from x in db.month_mapping
where x.PG_SUB_PROGRAM == SP
select x;
var PGRow = PGList.Select(x => new { x.PG }).Distinct().ToArray();
float[] PGContent = new float[12];
Dictionary<string,float[]> MonthRow = new Dictionary<string, float[]>();
foreach (var item in PGRow)
{
PGContent = new float[12];
var MonthList = from x in db.month_Web
where x.PG == item.PG
group x by new { x.ACCOUNTING_PERIOD, x.PG, x.Amount } into pggroup
select new { accounting_period = pggroup.Key.ACCOUNTING_PERIOD, amount = pggroup.Sum(x => x.Amount) };
foreach (var mon in MonthList)
{
int accounting_period = int.Parse(mon.accounting_period) - 1;
PGContent[accounting_period] = (float)mon.amount/1000000;
}
MonthRow[item.PG] = PGContent;
}
return Json(MonthRow, JsonRequestBehavior.AllowGet);
}
return View();
}
This code worked great for me since I am pulling from a Linq to SQL query instead of adding data directly into the code. My problems stemmed from mainly putting the data pulls outside of the foreach loops so it only pulled 1 piece of data from the SQL instead of all twelve months. I hope this helps some one else who is trying to pull data in from SQL data sources into multidimensional arrays.

Summing the previous values in an IEnumerable

I have a sequence of numbers:
var seq = new List<int> { 1, 3, 12, 19, 33 };
and I want to transform that into a new sequence where the number is added to the preceding numbers to create a new sequence:
{ 1, 3, 12, 19, 33 } --> {1, 4, 16, 35, 68 }
I came up with the following, but I dislike the state variable 'count'. I also dislike the fact that I'm using the values Enumerable without acting on it.
int count = 1;
var summed = values.Select(_ => values.Take(count++).Sum());
How else could it be done?

This is a common pattern in functional programming which in F# is called scan. It's like C#'s Enumerable.Aggregate and F#'s fold except that it yields the intermediate results of the accumulator along with the final result. We can implement scan in C# nicely with an extension method:
public static IEnumerable<U> Scan<T, U>(this IEnumerable<T> input, Func<U, T, U> next, U state) {
yield return state;
foreach(var item in input) {
state = next(state, item);
yield return state;
}
}
And then use it as follows:
var seq = new List<int> { 1, 3, 12, 19, 33 };
var transformed = seq.Scan(((state, item) => state + item), 0).Skip(1);

"Pure" LINQ:
var result = seq.Select((a, i) => seq.Take(i + 1).Sum());
One more "pure" LINQ O(n):
var res = Enumerable.Range(0, seq.Count)
.Select(a => a == 0 ? seq[a] : seq[a] += seq[a - 1]);
One more LINQ, with state maintenance:
var tmp = 0;
var result = les.Select(a => { tmp += a; return tmp; });

Stephen Swensen's answer is great, scan is exactly what you need. There is another version of scan though that doesn't require a seed, which would be slightly more appropriate for your exact problem.
This version requires that your output element type is the same as your input element type, which it is in your case, and gives the advantage of not needing you to pass in a 0 and then Skip the first (0) result.
You can implement this version of scan in C# as follows:
public static IEnumerable<T> Scan<T>(this IEnumerable<T> Input, Func<T, T, T> Accumulator)
{
using (IEnumerator<T> enumerator = Input.GetEnumerator())
{
if (!enumerator.MoveNext())
yield break;
T state = enumerator.Current;
yield return state;
while (enumerator.MoveNext())
{
state = Accumulator(state, enumerator.Current);
yield return state;
}
}
}
And then use it as follows:
IEnumerable<int> seq = new List<int> { 1, 3, 12, 19, 33 };
IEnumerable<int> transformed = seq.Scan((state, item) => state + item);

var seq = new List<int> { 1, 3, 12, 19, 33 };
var summed = new List<int>();
seq.ForEach(i => summed.Add(i + summed.LastOrDefault()));

Just to offer another alternative, albeit not really LINQ, you could write a yield-based function to do the aggregation:
public static IEnumerable<int> SumSoFar(this IEnumerable<int> values)
{
int sumSoFar = 0;
foreach (int value in values)
{
sumSoFar += value;
yield return sumSoFar;
}
}
Like BrokenGlass's this makes only a single pass over the data although unlike his returns an iterator not a list.
(Annoyingly you can't easily make this generic on the numeric type in the list.)

To use Linq and only iterate over the list once you could use a custom aggregator:
class Aggregator
{
public List<int> List { get; set; }
public int Sum { get; set; }
}
..
var seq = new List<int> { 1, 3, 12, 19, 33 };
var aggregator = new Aggregator{ List = new List<int>(), Sum = 0 };
var aggregatorResult = seq.Aggregate(aggregator, (a, number) => { a.Sum += number; a.List.Add(a.Sum); return a; });
var result = aggregatorResult.List;

var seq = new List<int> { 1, 3, 12, 19, 33 };
for (int i = 1; i < seq.Count; i++)
{
seq[i] += seq[i-1];
}

All answers already posted are working fine. I just would to add two things :
MoreLinq already have a Scan method in it, with a lot of very useful functions that are not native in LINQ. It's a great library to know for this kind of applications so I would like to recall it.
Before to see the answers above, I wrote my own Scan :
public static IEnumerable<T> Scan<T>(IEnumerable<T> input, Func<T, T, T> accumulator)
{
if (!input.Any())
yield break;
T state = input.First();
yield return state;
foreach (var value in input.Skip(1))
{
state = accumulator(state, value);
yield return state;
}
}
Works well but very probably less efficient than using #Nathan Phillips version due to multiple accesses to the collection. Enumerator probably does it better.