Be concise.
Be useful.
All contributions dictatorially edited by webmasters to match personal tastes.
Please do not paste any copyright violating material.
Please try to avoid dependencies to third-party libraries and frameworks.
(def y (->> x shuffle (take k)))allocate (sample(k))
do i=1,k
sample(i) = x(i)
end do
do i=k+1,n
call random_number(a)
j = 1 + int(i*a)
if (j .le. k) sample(j) = x(i)
end doThis is the R algorithm for reservoir sampling, see Wikipedia article.
randomSample :: Int -> [a] -> IO [a]
randomSample 0 x = pure []
randomSample k x = do
i <- randomRIO (0, length x - 1)
let (a, e:b) = splitAt i x
l <- randomSample (k-1) (a ++ b)
pure (e : l)Haskell doesn't allow non-deterministic functions. We have to use the IO type to let it know we'll only us this in an I/O context.
const idx = x.map((item, i) => i);
while (y.length < k) {
const i = parseInt(Math.random() * idx.length, 10);
y.push(x[[idx[i]]]);
idx.splice(i, 1);
}Note: lodash has a sample function.
Without native sampling in JS, create an array of unchosen indices of x and randomly pick them until y's length equals k.
Without native sampling in JS, create an array of unchosen indices of x and randomly pick them until y's length equals k.
List<T> y = new ArrayList<>(x);
shuffle(y);
y = y.subList(0, k);function RandArr(Max: Integer): TIntegerDynArray;
var
i, j, temp: Integer;
begin
SetLength(Result, Max+1);
for i := Low(Result) to High(Result) do Result[i] := i;
i := Length(Result);
while i > 0 do
begin
Dec(i);
j := RandomRange(0,i);
temp := Result[i];
Result[i] := Result[j];
Result[j] := temp;
end;
end;
var
Idx: TIntegerDynArray;
begin
Idx := RandArr(High(X));
SetLength(Y, k);
for i := 0 to k-1 do Y[i] := X[Idx];
end.First create an array (Idx) of random integers from 0 to number of elements in X minus 1.
It is created using the Sattolo algorithm.
Then create a list Y of k elements (all elements will be nil after creation).
Then copy k elements form X to Y using the first k elements of the Idx array to determine what element to pick from X
It is created using the Sattolo algorithm.
Then create a list Y of k elements (all elements will be nil after creation).
Then copy k elements form X to Y using the first k elements of the Idx array to determine what element to pick from X
y = x.sample(k)The original ordering is not preserved.
y := x shuffled first: k.Collections understand the message #take:, which returns k elements. For ordered collections however, this will simply default to taking the first k elements.