Why is there an extra variable in this javascript function?

Question

I'm a javascript newbie, and I can't figure out why the variable o exists in the following code:

var data = [ {"name":"alice", "group":1}, {"name":"bob", "group":2} ];

function getGroup(n) { return n.group; }

function network(nodes, index) {
    var gc = {};
    nodes.forEach(function(n) {
        var i = index(n), o;
        if (condition) {
            o = gc[i] || (gc[i] = {x:0,y:0,count:0});
            o.x += n.x;
            o.y += n.y;
            o.count += 1;
        }
    });
}

var net = network(nodes, getGroup)

It seems to me that the iterator in the network function would be better written this way:

function network(data, index) {
    var gc = {};
    nodes.forEach(function(n) {
        var i = index(n);
        if (condition) {
            gc[i] = gc[i] || (gc[i] = {x:0,y:0,count:0});
            gc[i].x += n.x;
            gc[i].y += n.y;
            gc[i].count += 1;
        }
    });
}

I think the second loop is equivalent and it eliminates a variable (and it's easier for me to read). Is the first version following some recommended practice in javascript? I can imagine that it would be good practice to have var o; inside the iterator to reduce its scope, but in this case o only refers to an object that's already referenced by gc[i]. Why not just eliminate it?

Please rephrase your question. I do not think it should contain "ANSWER", as that is the purpose of this site (Q&A). Also, rephrase it in a way that talks about caching and performance, as right now the title looks like a debug question. And why is conventions a tag? What about caching? — onebree, Jul 23 '15 at 20:28
I originally wanted some discussion about conventions but it didn't go like that; edits on the way — shortorian, Jul 23 '15 at 20:29
SO is not best about open-ended discussions for conventions. That would violate and lead to a flag for "too broad" or "primarily opinion-based." — onebree, Jul 23 '15 at 20:31
I get that open ended discussion is bad, but I don't think "is there a primary source for best practices like the python PEPs" (which I've now deleted) is an open ended question. — shortorian, Jul 23 '15 at 20:47
That may not be open-ended, but it runs up against another rule (#4 from [here](http://stackoverflow.com/help/on-topic)): _"Questions asking us to recommend or find a book, tool, software library, tutorial or other off-site resource are off-topic for Stack Overflow"_ Stick to asking about the programming problem itself. — Ted Hopp, Jul 23 '15 at 21:07
As far as primary sources, you can look through [my search query for JS style guides](https://duckduckgo.com/?q=javascript+style+guide+github&t=fedora&ia=software&iai=0). Mozilla Dev Network has great guides on web technologies. Also, great books are Eloquent JS. For other resrouces, check [here](https://github.com/Michael0x2a/curated-programming-resources/blob/master/resources.md) and [here](https://github.com/vhf/free-programming-books). — onebree, Jul 23 '15 at 21:11
You shouldn't edit your question with a new question. If you have a new question, you should post a new question. You've already accepted an answer, if you change the question the posted answers makes no sense. Also, you won't be able to pick a new answer for your new question. — Jan, Jul 23 '15 at 22:41
Jan, see comments above for the rationale for some edits. I can change the accepted answer whenever I want. By rolling back to the original question, you also removed some edits not related to your comment. — shortorian, Jul 23 '15 at 22:55
@dshort - you shouldn't change the question in such a way that invalidates the existing answers. If you have a follow up question ask it as a **new** question. — ChrisF, Jul 23 '15 at 23:07

Shripathi Kamath · Answer 1 · 2015-07-23T21:48:34.590

3

Likely because g[i] requires ~~indexing into an array~~ repeatedly looking up a property of an object, whereas storing the value in o caches the value once and provides for faster access subsequently.

edited Jul 23 '15 at 21:48

answered Jul 23 '15 at 01:32

Shripathi Kamath

310
3
10

Since you're storing the object in `gc[i]` and then storing the object reference in `o`, you're acting on the same object and it ends up being exactly the same as your code. Only more efficient. – Jan Jul 23 '15 at 01:35
Actually, `gc` is not an array, so this is not array index but simple property look-up. – Ted Hopp Jul 23 '15 at 02:14
Could you (or even @TedHopp) elaborate and correct that mistake in this answer? Also add some code formatting – onebree Jul 23 '15 at 20:29

score 1 · Accepted Answer · edited May 23 '17 at 12:14

This is a (minor) performance enhancement. It replaces three subscripting operations with a local variable assignment and three local variable accesses. I have my doubts whether the performance improvement would be noticeable. ~~See this thread for more info on the performance of various array operations.~~ (Actually, that info is somewhat irrelevant because gc is an object, not an array. These are plain old property look-ups being saved, not array indexing, so the performance increase might be a little more significant.)

score 1 · Answer 3 · answered Jul 23 '15 at 01:34

1

Even simpler

change

        gc[i] = gc[i] || (gc[i] = {x:0,y:0,count:0});

to

        gc[i] = gc[i] || {x:0,y:0,count:0};

answered Jul 23 '15 at 01:34

Jaromanda X

53,868
5
73
87

score 0 · Answer 4 · answered Jul 23 '15 at 01:35

0

Because it has to keep re-seeking the item in the array via the index. o is storing the reference to the location in memory so it's quicker to access later.

answered Jul 23 '15 at 01:35

wayofthefuture

8,339
7
36
53

Why is there an extra variable in this javascript function?

4 Answers4