Fast stable sorting algorithm implementation in javascript

Viewed 61885

I'm looking to sort an array of about 200-300 objects, sorting on a specific key and a given order (asc/desc). The order of results must be consistent and stable.

What would be the best algorithm to use, and could you provide an example of it's implementation in javascript?

Thanks!

16 Answers

It is possible to get a stable sorting from a non-stable sort function.

Before sorting you get the position of all the elements. In your sort condition, if both elements are equal, then you sort by the position.

Tada! You've got a stable sort.

I've written an article about it on my blog if you want to know more about this technique and how to implement it: http://blog.vjeux.com/2010/javascript/javascript-sorting-table.html

Somewhat shorter version of the same thing using ES2017 features like arrow functions and destructuring:

Function

var stableSort = (arr, compare) => arr
  .map((item, index) => ({item, index}))
  .sort((a, b) => compare(a.item, b.item) || a.index - b.index)
  .map(({item}) => item)

It accepts input array and compare function:

stableSort([5,6,3,2,1], (a, b) => a - b)

It also returns new array instead of making in-place sort like the built-in Array.sort() function.

Test

If we take the following input array, initially sorted by weight:

// sorted by weight
var input = [
  { height: 100, weight: 80 },
  { height: 90, weight: 90 },
  { height: 70, weight: 95 },
  { height: 100, weight: 100 },
  { height: 80, weight: 110 },
  { height: 110, weight: 115 },
  { height: 100, weight: 120 },
  { height: 70, weight: 125 },
  { height: 70, weight: 130 },
  { height: 100, weight: 135 },
  { height: 75, weight: 140 },
  { height: 70, weight: 140 }
]

Then sort it by height using stableSort:

stableSort(input, (a, b) => a.height - b.height)

Results in:

// Items with the same height are still sorted by weight 
// which means they preserved their relative order.
var stable = [
  { height: 70, weight: 95 },
  { height: 70, weight: 125 },
  { height: 70, weight: 130 },
  { height: 70, weight: 140 },
  { height: 75, weight: 140 },
  { height: 80, weight: 110 },
  { height: 90, weight: 90 },
  { height: 100, weight: 80 },
  { height: 100, weight: 100 },
  { height: 100, weight: 120 },
  { height: 100, weight: 135 },
  { height: 110, weight: 115 }
]

However sorting the same input array using the built-in Array.sort() (in Chrome/NodeJS):

input.sort((a, b) => a.height - b.height)

Returns:

var unstable = [
  { height: 70, weight: 140 },
  { height: 70, weight: 95 },
  { height: 70, weight: 125 },
  { height: 70, weight: 130 },
  { height: 75, weight: 140 },
  { height: 80, weight: 110 },
  { height: 90, weight: 90 },
  { height: 100, weight: 100 },
  { height: 100, weight: 80 },
  { height: 100, weight: 135 },
  { height: 100, weight: 120 },
  { height: 110, weight: 115 }
]

Resources

Update

Array.prototype.sort is now stable in V8 v7.0 / Chrome 70!

Previously, V8 used an unstable QuickSort for arrays with more than 10 elements. Now, we use the stable TimSort algorithm.

source

I know this has been plenty answered. I just wanted to go ahead an post a quick TS implementation for anyone that landed here looking for that.

export function stableSort<T>( array: T[], compareFn: ( a: T, b: T ) => number ): T[] {
    const indices = array.map( ( x: T, i: number ) => ( { element: x, index: i } ) );

    return indices.sort( ( a, b ) => {
        const order = compareFn( a.element, b.element );
        return order === 0 ? a.index - b.index : order;
    } ).map( x => x.element );
}

The method does no longer run in-place, as the native sort does. I also want to point out that it is not the most efficient. It adds two loops of the order O(n). though sort itself is most likely O(n log(n)) so it's less than that.

Some of the solutions mentioned are more performant, thought this might be less code, also using internal Array.prototype.sort.

(For a Javascript solution, just remove all the types)

According to the v8 dev blog and caniuse.com Array.sort is already stable as required by the spec in modern browsers, so you don't need to roll your own solution. The only exception I can see is Edge, which should soon move over to chromium and support it as well.

function sort(data){
    var result=[];
    var array = data;
    const array2=data;
    const len=array2.length;
    for(var i=0;i<=len-1;i++){
    var min = Math.min.apply(Math,array)
    result.push(min);
    var index=array.indexOf(min)
    array.splice(index,1);
    }
    return result;
}   
sort([9,8,5,7,9,3,9,243,4,5,6,3,4,2,4,7,4,9,55,66,33,66]);

Related