数组连接与字符串连接

小开

I can definitely say that using Array.join() is faster. I've worked on a few pieces of JavaScript code and sped up performance significantly by removing string manipulation in favor of arrays.

小开

最佳答案

String concatenation is faster in ECMAScript. Here's a benchmark I created to show you:

http://jsben.ch/#/OJ3vo

小开

According to this Google document titled 'Optimizing JavaScript code' string concat is slower then array join but apparently this is not true for modern Javascript engines.

I made a benchmark for the Fibonacci test example that they used in the document and it shows that concatenating (gluing) the string is almost 4x as fast as using Array join.

小开

From 2011 and into the modern day ...

See the following join rewrite using string concatenation, and how much slower it is than the standard implementation.

// Number of times the standard `join` is faster, by Node.js versions:
// 0.10.44: ~2.0
// 0.11.16: ~4.6
// 0.12.13: ~4.7
// 4.4.4: ~4.66
// 5.11.0: ~4.75
// 6.1.0: Negative ~1.2 (something is wrong with 6.x at the moment)
function join(sep) {
var res = '';
if (this.length) {
res += this[0];
for (var i = 1; i < this.length; i++) {
res += sep + this[i];
}
}
return res;
}

The moral is - do not concatenate strings manually, always use the standard join.

小开

The spread operator, written with three consecutive dots ( ... ), is new in ES6 and gives you the ability to expand, or spread, iterable objects into multiple elements.

const books = ["Don Quixote", "The Hobbit", "Alice in Wonderland", "Tale of Two Cities"];
console.log(...books);

Prints: Don Quixote The Hobbit Alice in Wonderland Tale of Two Cities

小开

Manual concatenation is faster, for a numeric array of fixed length.

Here's a JSPerf test that tests these two operations:

zxy.join('/')


// versus


zxy[0] + '/' + zxy[1] + '/' + zxy[2]


// given the array


zxy = [1, 2, 3]


// resulting in the string '0/1/2'

Results: Using Chrome 64.0.3282.186, Array.join was 46% slower.

小开

It depends:

Chromium 79.0.3945

Array Join is 30% slower

Firefox 71.0.0

String Concat is 90% slower

https://jsperf.com/lin-array-join-vs-string-concat

小开

join is way faster when the array of strings already exists. The real comparison would be comparing:

push elements into the array and then join them to build the string
concatenate string every time without using the array

For small number of iteration and strings, it does not matter whether you use push-and-join or concatenate. However, for large number of strings, array push and join seems to be faster in both chrome and firefox.

Here is the code and the test results for 10 to 10 million strings:

Chrome:

strings 10
join-only: 0.01171875 ms
push-join: 0.137939453125 ms
concatenate: 0.01513671875 ms
strings 100
join-only: 0.01416015625 ms
push-join: 0.13427734375 ms
concatenate: 0.0830078125 ms
strings 1000
join-only: 0.048095703125 ms
push-join: 0.47216796875 ms
concatenate: 0.5517578125 ms
strings 10000
join-only: 0.465087890625 ms
push-join: 5.47314453125 ms
concatenate: 4.9619140625 ms
strings 100000
join-only: 7.6240234375 ms
push-join: 57.37109375 ms
concatenate: 67.028076171875 ms
strings 1000000
join-only: 67.666259765625 ms
push-join: 319.3837890625 ms
concatenate: 609.8369140625 ms
strings 10000000
join-only: 824.260009765625 ms
push-join: 3207.129150390625 ms
concatenate: 5959.56689453125 ms

Firefox:

strings 10
join-only: 0ms
push-join: 1ms
concatenate: 0ms
strings 100
join-only: 0ms
push-join: 0ms
concatenate: 0ms
strings 1000
join-only: 0ms
push-join: 1ms
concatenate: 0ms
strings 10000
join-only: 1ms
push-join: 2ms
concatenate: 0ms
strings 100000
join-only: 5ms
push-join: 11ms
concatenate: 8ms
strings 1000000
join-only: 39ms
push-join: 88ms
concatenate: 98ms
strings 10000000
join-only: 612ms
push-join: 1095ms
concatenate: 3249ms

Code to test:

for (var n = 10; n <= 10000000; n*=10) {
    

var iterations = n;


console.log("strings", iterations);
console.time("push-join");
arr = [];
for (var i = 0; i< iterations; i++) {
arr.push("a b c d e f g h i j k l m");
}
console.time("join-only");
content = arr.join(",");
console.timeEnd("join-only");
console.timeEnd("push-join");


content = "";


console.time("concatenate");
for (var i = 0; i< iterations; i++) {
content += "a b c d e f g h i j k l m";
}
console.timeEnd("concatenate");


}

小开

2021 test

See code below. Results:

Firefox: push+join is 80% slower than string concat, in regular usage.

Chrome: push+join is 140% slower than string concat, in regular usage.

function test(items = 100, rep = 1000000) {
let str


console.time('concat')
for (let r = 0; r < rep; r++) {
str = ''
for (let i = 0; i < items; i++) {
str += i
}
}
console.timeEnd('concat')


console.time('push+join')
for (let r = 0; r < rep; r++) {
const arr = []
for (let i = 0; i < items; i++) {
arr.push(i)
}
str = arr.join('')
}
console.timeEnd('push+join')
}

小开

i think it is not only performance problem, but memory too

as string is immutable, means every time you concat a string, a extra string create in your memory

but array is mutable, mean it will keep the same memory address

of course it is also depend on languages, in general , array is a better solution