Saturday 27 August 2016

Fastest way to duplicate an array in JavaScript - slice vs. 'for' loop



In order to duplicate an array in JavaScript: which of the following is faster to use?



Slice method




var dup_array = original_array.slice();


For loop



for(var i = 0, len = original_array.length; i < len; ++i)
dup_array[i] = original_array[i];






I know both ways do only a shallow copy: if original_array contains references to objects, objects won't be cloned, but only the references will be copied, and therefore both arrays will have references to the same objects.
But this is not the point of this question.



I'm asking only about speed.


Answer



There are at least 5 (!) ways to clone an array:





  • loop

  • slice

  • Array.from()

  • concat

  • spread operator (FASTEST)



There has been a huuuge BENCHMARKS thread, providing following information:





  • for blink browsers slice() is the fastest method, concat() is a bit slower, and while loop is 2.4x slower.


  • for other browsers while loop is the fastest method, since those browsers don't have internal optimizations for slice and concat.




This remains true in Jul 2016.



Below are simple scripts that you can copy-paste into your browser's console and run several times to see the picture. They output milliseconds, lower is better.



while loop




n = 1000*1000;
start = + new Date();
a = Array(n);
b = Array(n);
i = a.length;
while(i--) b[i] = a[i];
console.log(new Date() - start);


slice




n = 1000*1000;
start = + new Date();
a = Array(n);
b = a.slice();
console.log(new Date() - start);


Please note that these methods will clone the Array object itself, array contents however are copied by reference and are not deep cloned.




origAr == clonedArr //returns false
origAr[0] == clonedArr[0] //returns true

No comments:

Post a Comment

c++ - Does curly brackets matter for empty constructor?

Those brackets declare an empty, inline constructor. In that case, with them, the constructor does exist, it merely does nothing more than t...