问题描述
从 Javascript 1.7 开始,有一个 Iterator 对象,它允许这样做:
Since Javascript 1.7 there is an Iterator object, which allows this:
var a={a:1,b:2,c:3};
var it=Iterator(a);
function iterate(){
try {
console.log(it.next());
setTimeout(iterate,1000);
}catch (err if err instanceof StopIteration) {
console.log("End of record.
");
} catch (err) {
console.log("Unknown error: " + err.description + "
");
}
}
iterate();
node.js 中有这样的东西吗?
is there something like this in node.js ?
我现在正在使用:
function Iterator(o){
/*var k=[];
for(var i in o){
k.push(i);
}*/
var k=Object.keys(o);
return {
next:function(){
return k.shift();
}
};
}
但是通过将所有对象键存储在 k
中会产生大量开销.
but that produces a lot of overhead by storing all the object keys in k
.
推荐答案
你想要的是对对象或数组的惰性迭代.这在 ES5 中是不可能的(因此在 node.js 中是不可能的).我们最终会得到这个.
What you want is lazy iteration over an object or array. This is not possible in ES5 (thus not possible in node.js). We will get this eventually.
唯一的解决方案是找到一个扩展 V8 的节点模块来实现迭代器(可能还有生成器).我找不到任何实现.您可以查看 spidermonkey 源代码并尝试使用 C++ 将其编写为 V8 扩展.
The only solution is finding a node module that extends V8 to implement iterators (and probably generators). I couldn't find any implementation. You can look at the spidermonkey source code and try writing it in C++ as a V8 extension.
您可以尝试以下方法,但它也会将所有键加载到内存中
You could try the following, however it will also load all the keys into memory
Object.keys(o).forEach(function(key) {
var val = o[key];
logic();
});
然而,由于 Object.keys
是一种原生方法,它可能会带来更好的优化.
However since Object.keys
is a native method it may allow for better optimisation.
基准测试
如您所见,Object.keys 明显更快.实际的内存存储是否更优化是另一回事.
As you can see Object.keys is significantly faster. Whether the actual memory storage is more optimum is a different matter.
var async = {};
async.forEach = function(o, cb) {
var counter = 0,
keys = Object.keys(o),
len = keys.length;
var next = function() {
if (counter < len) cb(o[keys[counter++]], next);
};
next();
};
async.forEach(obj, function(val, next) {
// do things
setTimeout(next, 100);
});
这篇关于迭代 node.js 中的对象键的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持跟版网!