在mongo中执行优先级查询答案

【问题标题】：performing priority query in mongo在mongo中执行优先级查询
【发布时间】：2016-04-16 21:48:24
【问题描述】：

示例文档：

{"name":"John", "age":35, "address":".....",.....}

join_month=3 优先级为 1 的员工
地址中包含字符串“Avenue”的员工优先级为 2
地址包含字符串“Street”的员工优先级为 3
地址中包含字符串“Road”的员工优先级为4

到目前为止，我正处于这个阶段：

db.collection.aggregate([
    { "$match": { 
        "$or": [ 
            { "join_month": 3 }, 
            { "address": /.*Avenue.*/i }, 
            { "address": /.*Street.*/i }, 
            { "address": /.*Road.*/i }
        ] 
    }}, 
    { "$project": { 
        "name": 1, 
        "age": 1,
        "_id": 0, 
        "priority": { ?????????? } 
    }}, 
    { "$sort":  { "priority": 1 } }
])

我被困在优先领域。我应该放什么？

【问题讨论】：

标签： regex mongodb mapreduce mongodb-query

【解决方案1】：

使用聚合框架，您“理想地”希望在 $project 管道步骤，但不幸的是 MongoDB 尚未支持这一点。目前开放的有一张 JIRA 票证$project filter using $regex

但是，一种解决方法（虽然不是性能方面的最佳解决方案）是使用 map-reduce。考虑填充测试集合：

db.test.insert([
    { _id: 0, "join_month": 12, "address": "33 Point Avenue", "name": "John", "age":35 },
    { _id: 1, "join_month": 10, "address": "2A Broad Street, Surbub", "name": "Jane", "age":21 },
    { _id: 2, "join_month": 3, "address": "127 Umpstreeten Road, Surbub", "name": "Alan", "age":63 },
    { _id: 3, "join_month": 3, "address": "127 Umpstreeten Road, Surbub", "name": "Louise", "age":30 }
])

将地图函数定义为：

var mapper = function() {
    var priority;
    if (this.join_month==3){
        priority = 1;
    }
    else if (this.address.match(/Avenue/i)){
        priority = 2;
    }
    else if (this.address.match(/Street/i)){
        priority = 3;
    }
    else if (this.address.match(/Road/i)){
        priority = 4;
    }
    else {
        priority = 99;
    }

    var value = {
        "name": this.name, 
        "age": this.age,
        "priority": priority
    };
    emit( this._id, value );        
};

reduce 函数如下：

var reducer = function() { };

然后在测试集合上运行mapduce操作，并将结果存入集合mr_result

db.test.mapReduce(mapper, reducer, {
    "out": 'mr_result'
    "query": {
        "$or": [ 
            { "join_month": 3 }, 
            { "address": /.*Avenue.*/i }, 
            { "address": /.*Street.*/i }, 
            { "address": /.*Road.*/i }
        ] 
    }
})

查询结果集合：

db.mr_result.find().sort({ "priority": 1})

样本输出

{ "_id" : 2, "value" : { "name" : "Alan", "age" : 63, "priority" : 1 } }
{ "_id" : 3, "value" : { "name" : "Louise", "age" : 30, "priority" : 1 } }
{ "_id" : 0, "value" : { "name" : "John", "age" : 35, "priority" : 2 } }
{ "_id" : 1, "value" : { "name" : "Jane", "age" : 21, "priority" : 3 } }

【讨论】：