0

I've been learning Ramda and wanted to know how to sum n-arrays by index. Below is what I was able to do with 2 arrays. How can I make this method scale?

i.e. I'd like to be able to do this: sumByIndex( arr1, arr2, ..., arrn )

Given lists x and y, the resultant array should yield [x0 + y0, x1 + y1, ..., xn + yn]. So for the case of n-array, the resultant array should be [ a[0][0] + a[1][0] + ... a[n][0], a[0][1] + a[1][1] + ... a[n][1], ..., a[0][n] + a[1][n] + ... + a[n][n] ] where a[n] is an array as an argument at position n.

var array1 = [1,2,3];
var array2 = [2,4,6];

var sumByIndex = R.map(R.sum);
var result = sumByIndex(R.zip(array1, array2));

$('pre').text(JSON.stringify(result, true));
<script src="https://ajax.googleapis.com/ajax/libs/jquery/2.1.1/jquery.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/ramda/0.18.0/ramda.min.js"></script>
<pre></pre>
Pete
  • 3,246
  • 4
  • 24
  • 43
  • What is the expected result? `[3, 6, 9]`? – thefourtheye Nov 18 '15 at 03:39
  • @thefourtheye yes the expected result is `[3, 6, 9]`. I added more information for clarity... I see my mistake now: `sumArrays` is a misleading name. I've renamed it to `sumByIndex`. – Pete Nov 18 '15 at 03:47

3 Answers3

1

To achieve this, we'll start by creating a few generic helper functions:

// a new version of `map` that includes the index of each item
var mapI = R.addIndex(R.map);

// a function that can summarise a list of lists by their respective indices
var zipNReduce = R.curry(function(fn, lists) {
  return mapIndexed(function (_, n) {
    return fn(R.pluck(n, lists));
  }, R.head(lists));
});

Once we have these, we can create sumByIndex, by passing R.sum to the zipNReduce defined above.

var sumByIndex = zipNReduce(R.sum);
sumByIndex([[1, 2, 3], [4, 5, 6], [7, 8, 9]]); // [12, 15, 18]

If you'd prefer to create a function that accepts a varying number of arrays as arguments rather than the array of arrays, you can simply wrap it with R.unapply:

var sumByIndex_ = R.unapply(sumByIndex);
sumByIndex_([1, 2, 3], [4, 5, 6], [7, 8, 9]); // [12, 15, 18]

And if you're potentially dealing with lists of different sizes, we can swap out R.sum with a slight variation which defaults undefined values to zero:

var sumDefaultZero = R.reduce(R.useWith(R.add, [R.identity, R.defaultTo(0)]), 0);
var sumByIndexSafe = zipNReduce(sumDefaultZero);
sumByIndexSafe([[1, 2, 3], [], [7, 9]]); // [8, 11, 3]
Scott Christopher
  • 6,458
  • 23
  • 26
1

I found the answer a little verbose. Better keep it simple.

import { compose, map, unnest, zip, sum } from 'ramda';

const a = [1,2,3]
const b = [4,5,6]
const c = [7,8,9]

function groupByIndex(/*[1,2,4], [4,5,6], ...*/) {
  return [...arguments].reduce(compose(map(unnest), zip));
}

const sumByIndex = map(sum);
const res = sumByIndex(groupByIndex(a,b,c))
// => [12,15,18]
ja0nz
  • 1,891
  • 1
  • 9
  • 5
1

I am bit late to the party but assuming that all arrays are of the same length, we can take the first array as the initial value of a reducing function. The rest is iterated over with a zipWith function that adds two numbers.

const {unapply, converge, reduce, zipWith, add, head, tail} = R;

const a = [1,2,3];
const b = [4,5,6];
const c = [7,8,9];

var zipSum = 
  unapply(
    converge(
      reduce(zipWith(add)), [
        head,
        tail]));


var res = zipSum(a, b, c);

console.log(res);
<script src="https://cdnjs.cloudflare.com/ajax/libs/ramda/0.25.0/ramda.js"></script>
customcommander
  • 17,580
  • 5
  • 58
  • 84