另一种可以同时处理任意数量数组的索引方法:

// Calculate intersection of multiple array or object values.
function intersect (arrList) {
    var arrLength = Object.keys(arrList).length;
        // (Also accepts regular objects as input)
    var index = {};
    for (var i in arrList) {
        for (var j in arrList[i]) {
            var v = arrList[i][j];
            if (index[v] === undefined) index[v] = 0;
            index[v]++;
        };
    };
    var retv = [];
    for (var i in index) {
        if (index[i] == arrLength) retv.push(i);
    };
    return retv;
};

它只适用于可以作为字符串计算的值，你应该将它们作为一个数组传递:

intersect ([arr1, arr2, arr3...]);

.．.但它透明地接受对象作为参数或任何要交叉的元素(总是返回公共值的数组)。例子:

intersect ({foo: [1, 2, 3, 4], bar: {a: 2, j:4}}); // [2, 4]
intersect ([{x: "hello", y: "world"}, ["hello", "user"]]); // ["hello"]

编辑:我只是注意到，这是，在某种程度上，有点bug。

也就是说:我在编码时认为输入数组本身不能包含重复(正如所提供的示例那样)。

但如果输入数组恰好包含重复，就会产生错误的结果。示例(使用下面的实现):

intersect ([[1, 3, 4, 6, 3], [1, 8, 99]]);
// Expected: [ '1' ]
// Actual: [ '1', '3' ]

幸运的是，这很容易通过添加二级索引来解决。那就是:

变化:

        if (index[v] === undefined) index[v] = 0;
        index[v]++;

by:

        if (index[v] === undefined) index[v] = {};
        index[v][i] = true; // Mark as present in i input.

,:

         if (index[i] == arrLength) retv.push(i);

by:

         if (Object.keys(index[i]).length == arrLength) retv.push(i);

完整的例子:

// Calculate intersection of multiple array or object values.
function intersect (arrList) {
    var arrLength = Object.keys(arrList).length;
        // (Also accepts regular objects as input)
    var index = {};
    for (var i in arrList) {
        for (var j in arrList[i]) {
            var v = arrList[i][j];
            if (index[v] === undefined) index[v] = {};
            index[v][i] = true; // Mark as present in i input.
        };
    };
    var retv = [];
    for (var i in index) {
        if (Object.keys(index[i]).length == arrLength) retv.push(i);
    };
    return retv;
};

intersect ([[1, 3, 4, 6, 3], [1, 8, 99]]); // [ '1' ]

基于Anon的出色回答，这个函数返回两个或多个数组的交集。

function arrayIntersect(arrayOfArrays)
{        
    var arrayCopy = arrayOfArrays.slice(),
        baseArray = arrayCopy.pop();

    return baseArray.filter(function(item) {
        return arrayCopy.every(function(itemList) {
            return itemList.indexOf(item) !== -1;
        });
    });
}

这是一个提议的标准:对于当前阶段2的提议https://github.com/tc39/proposal-set-methods，您可以使用

mySet.intersection(mySet2);

在此之前，你可以使用Immutable.js的Set，它激发了这个提议

Immutable.Set(mySet).intersect(mySet2)

希望这有助于所有版本。

function diffArray(arr1, arr2) {
  var newArr = [];

  var large = arr1.length>=arr2.length?arr1:arr2;
  var small = JSON.stringify(large) == JSON.stringify(arr1)?arr2:arr1;
  for(var i=0;i<large.length;i++){
    var copyExists = false; 
    for(var j =0;j<small.length;j++){
      if(large[i]==small[j]){
        copyExists= true;
        break;
      }
    }
    if(!copyExists)
      {
        newArr.push(large[i]);
      }
  }

  for(var i=0;i<small.length;i++){
    var copyExists = false; 
    for(var j =0;j<large.length;j++){
      if(large[j]==small[i]){
        copyExists= true;
        break;
      }
    }
    if(!copyExists)
      {
        newArr.push(small[i]);
      }
  }


  return newArr;
}

//在线性时间内返回数组a中也在b中的元素: 函数相交(a, b) { 返回a.filter (Set.prototype。new Set(b)); ｝ / /例如: console.log(相交([1,2,3],[2、3、4、5]));

我推荐上述简洁的解决方案，它在大输入上优于其他实现。如果在小输入上的性能很重要，请检查下面的替代方案。

备选方案和性能比较:

有关替代实现，请参阅下面的代码片段，并检查https://jsperf.com/array-intersection-comparison以进行性能比较。

function intersect_for(a, b) { const result = []; const alen = a.length; const blen = b.length; for (let i = 0; i < alen; ++i) { const ai = a[i]; for (let j = 0; j < blen; ++j) { if (ai === b[j]) { result.push(ai); break; } } } return result; } function intersect_filter_indexOf(a, b) { return a.filter(el => b.indexOf(el) !== -1); } function intersect_filter_in(a, b) { const map = b.reduce((map, el) => {map[el] = true; return map}, {}); return a.filter(el => el in map); } function intersect_for_in(a, b) { const result = []; const map = {}; for (let i = 0, length = b.length; i < length; ++i) { map[b[i]] = true; } for (let i = 0, length = a.length; i < length; ++i) { if (a[i] in map) result.push(a[i]); } return result; } function intersect_filter_includes(a, b) { return a.filter(el => b.includes(el)); } function intersect_filter_has_this(a, b) { return a.filter(Set.prototype.has, new Set(b)); } function intersect_filter_has_arrow(a, b) { const set = new Set(b); return a.filter(el => set.has(el)); } function intersect_for_has(a, b) { const result = []; const set = new Set(b); for (let i = 0, length = a.length; i < length; ++i) { if (set.has(a[i])) result.push(a[i]); } return result; }

Firefox 53的结果:

Ops/sec on large arrays (10,000 elements): filter + has (this) 523 (this answer) for + has 482 for-loop + in 279 filter + in 242 for-loops 24 filter + includes 14 filter + indexOf 10 Ops/sec on small arrays (100 elements): for-loop + in 384,426 filter + in 192,066 for-loops 159,137 filter + includes 104,068 filter + indexOf 71,598 filter + has (this) 43,531 (this answer) filter + has (arrow function) 35,588

最简单、最快的O(n)和最短的方式:

函数交点(a, b) { const setA = new Set(a); 返回b.filter(value => setA.has(value)); ｝ console.log(十字路口([1,2,3],[2、3、4、5)))

@nbarbosa几乎有相同的答案，但他将两个数组强制转换为Set，然后再转换回数组。不需要任何额外的选角。