Consider the following simple array:
var foods = ['hotdog', 'hamburger', 'soup', 'sandwich', 'hotdog', 'watermelon', 'hotdog'];
With underscore
, is there a function or combination of functions I can use to select the most frequently occurring value (in this case it's hotdog
)?
var foods = ['hotdog', 'hamburger', 'soup', 'sandwich', 'hotdog', 'watermelon', 'hotdog'];
var result = _.chain(foods).countBy().pairs().max(_.last).head().value();
console.log(result);
<script src="https://cdnjs.cloudflare.com/ajax/libs/underscore.js/1.8.3/underscore.js"></script>
countBy:
Sorts a list into groups and returns a count for the number of objects
in each group.
pairs:
Convert an object into a list of [key, value]
pairs.
max:
Returns the maximum value in list. If an iterator function is provided, it will be used on each value to generate the criterion by which the value is ranked.
last:
Returns the last element of an array
head:
Returns the first element of an array
chain:
Returns a wrapped object. Calling methods on this object will continue to return wrapped objects until value is used.
value:
Extracts the value of a wrapped object.
You can do this in one pass using _.reduce
. The basic idea is to keep track of the word frequencies and the most common word at the same time:
var o = _(foods).reduce(function(o, s) {
o.freq[s] = (o.freq[s] || 0) + 1;
if(!o.freq[o.most] || o.freq[s] > o.freq[o.most])
o.most = s;
return o;
}, { freq: { }, most: '' });
That leaves 'hotdot'
in o.most
.
Demo: http://jsfiddle.net/ambiguous/G9W4m/
You can also do it with each
(or even a simple for
loop) if you don't mind predeclaring the cache variable:
var o = { freq: { }, most: '' };
_(foods).each(function(s) {
o.freq[s] = (o.freq[s] || 0) + 1;
if(!o.freq[o.most] || o.freq[s] > o.freq[o.most])
o.most = s;
});
Demo: http://jsfiddle.net/ambiguous/WvXEV/
You could also break o
into two pieces and use a slightly modified version of the above, then you wouldn't have to say o.most
to get 'hotdog'
.