按相关性能问题排序

Question

是否有一个函数/方法在mongoose(和或mongodb)可用于排序查询结果基于相关性(匹配查询参数的最高数量)?

下面的例子是我目前使用的(查询使用$in:[]，其他方面相同)-我的集合相当小，所以性能很好，但在较大的集合上，它会显着减慢速度。

或者，如果有一个更好的执行方法(mongoose/mongodb之外)，我很乐意知道它。

的例子:

var docs = [
    {
        fruits: ['apple', 'orange', 'tomato'],
        colors: ['blue', 'green'],
        // relevance: 3
    },
    {
        fruits: ['apple', 'carrot'],
        colors: ['red', 'green'],
        // relevance: 2
    }
]
var query = {fruits: ['apple', 'orange'], colors: ['green']}
docs.forEach(function(doc){
    var relevance = 0
    Object.keys(query).forEach(function(_query){
        var arrays = [doc[_query], query[_query]]
        var result = arrays.shift().filter(function(v) {
            return arrays.every(function(a) {
                return a.indexOf(v) !== -1;
            });
        });
        relevance += result.length
    })
    doc.relevance = relevance
})

结果:

var docs = [
    {
        fruits: ['apple', 'orange', 'tomato'],
        colors: ['blue', 'green'],
        relevance: 3
    },
    {
        fruits: ['apple', 'carrot'],
        colors: ['red', 'green'],
        relevance: 2
    }
]

Answer 1

您可以使用聚合:

db.getCollection('docs').aggregate([
{$match: {fruits: {$in: ['apple', 'orange']}, colors: {$in: ['green']}}},
{$project: {
    relevance: {
        $sum: [
          {$cond: {if: { "$setIsSubset": [['orange'], "$fruits" ]}, then: 1, else: 0}},
          {$cond: {if: { "$setIsSubset": [['apple'], "$fruits" ]}, then: 1, else: 0}},
          {$cond: {if: { "$setIsSubset": [['green'], "$colors" ]}, then: 1, else: 0}}]
    },
    doc: '$$ROOT'}}
])

结果:

/* 1 */
{
    "_id" : ObjectId("57be8a9b65d2835e960df543"),
    "relevance" : 3,
    "doc" : {
        "_id" : ObjectId("57be8a9b65d2835e960df543"),
        "fruits" : [ 
            "apple", 
            "orange", 
            "tomato"
        ],
        "colors" : [ 
            "blue", 
            "green"
        ]
    }
}
/* 2 */
{
    "_id" : ObjectId("57be8aa865d2835e960df544"),
    "relevance" : 2,
    "doc" : {
        "_id" : ObjectId("57be8aa865d2835e960df544"),
        "fruits" : [ 
            "apple", 
            "carrot"
        ],
        "colors" : [ 
            "red", 
            "green"
        ]
    }
}