mongodb--reduce并行处理框架

reduce 命令

db.runCommand(
               {
                 mapReduce: <collection>,
                 map: <function>,
                 reduce: <function>,
                 finalize: <function>,
                 out: <output>,
                 query: <document>,
                 sort: <document>,
                 limit: <number>,
                 scope: <document>,
                 jsMode: <boolean>,
                 verbose: <boolean>,
                 bypassDocumentValidation: <boolean>
               }
             )

1. mapReduce 你要做并行处理的集合名字
2. map： 分组依据， emit(key,value) => key => group的key，value => 是你key的数组 key=25, values=【“jack","mary"]
3. reduce: function(key,values) 里面有两个参数， 第一个参数是key：group操作的key，第二个参数是values： 我们刚才说到的 【数组】，类似group中的finalize函数。
4. finalize： 这个呢，还是和group中的 finalize 一样，会对每组的结构进行一个最后修改。也可以不修改。
5. out： 就是我们进行mapreduce之后，将结果存放的位置，一般是“集合类型“。
6. query: 这个很简单，和group的query一样
7. sort： 一样的到底，对map操作之前的文档进行排序。
8. limit： 在map之前限制文档的个数
9. scope： 给reduce,query之类的函数提供变量访问。
10. 我们常用到的是： mapReduce， map，reduce， out，query，

二：我们看一下注意事项

我们还是用一个person集合，其中有name和age。

上一节课程我们将group的时候，做了一个对age进行分组的一个例子

｛age:26,count:2｝
｛age:29,count:2｝
｛age:31,count:2｝

1. 第一步：插入数据源

> db.person.find()
{ "_id" : ObjectId("5678a3fc763ea69ae9780c62"), "name" : "mary", "age" : 26 }
{ "_id" : ObjectId("5678a403763ea69ae9780c63"), "name" : "jackson", "age" : 26 }
{ "_id" : ObjectId("5678a40a763ea69ae9780c64"), "name" : "john", "age" : 29 }
{ "_id" : ObjectId("5678a410763ea69ae9780c65"), "name" : "hxc", "age" : 31 }
{ "_id" : ObjectId("5678a416763ea69ae9780c66"), "name" : "mongodb", "age" : 31 }
{ "_id" : ObjectId("5678a41d763ea69ae9780c67"), "name" : "redis", "age" : 29 }

1. 执行命令

 db.runCommand({
  mapreduce:"person",
  map:function(){
      emit(this.age,this.name);		 //相当于group的key
   },
   reduce:function(key,values){	      //{key:26,values:["mary","jackson"]}
      return {"key":key,"values":values};
   },
   finalize:function(key,values){
      return {"age":key,"count":values.length};
   },
   out:"t"    //最终的存放结果
 })

db.runCommand({
  mapreduce:"person",
  map:function(){
      emit(this.age,this.name);
   },
   reduce:function(key,values){
      return {"key":key,"values":values};
   },
  out:"t"
 })

db.runCommand({
     mapreduce:"person",
     map:function(){
        emit(this.age,this.name)
     },
     reduce:function(key,values){
        return values.length;
      },
      out:"t",
     finalize:function(key,values){
       return values;
     },
     query:{"$where":function(){return this.age>28}}
})
db.person.find()
{ "_id" : ObjectId("5678a3fc763ea69ae9780c62"), "name" : "mary", "age" : 26 }
{ "_id" : ObjectId("5678a403763ea69ae9780c63"), "name" : "jackson", "age" : 26 }
{ "_id" : ObjectId("5678a40a763ea69ae9780c64"), "name" : "john", "age" : 29 }
{ "_id" : ObjectId("5678a410763ea69ae9780c65"), "name" : "hxc", "age" : 31 }
{ "_id" : ObjectId("5678a416763ea69ae9780c66"), "name" : "mongodb", "age" : 31 }
{ "_id" : ObjectId("5678a41d763ea69ae9780c67"), "name" : "redis", "age" : 29 }

db.runCommand({
...      mapreduce:"person",
...      map:function(){
...         emit(this.age,this.name)
...      },
...      reduce:function(key,values){
...         return {"key":key,"values":values}
...       },
...       out:"t"
... })
{
	"result" : "t",
	"timeMillis" : 20,
	"counts" : {
		"input" : 6,
		"emit" : 6,
		"reduce" : 3,
		"output" : 3
	},
	"ok" : 1
}
 db.t.find()
{ "_id" : 26, "value" : { "key" : 26, "values" : [  "mary",  "jackson" ] } }
{ "_id" : 29, "value" : { "key" : 29, "values" : [  "john",  "redis" ] } }
{ "_id" : 31, "value" : { "key" : 31, "values" : [  "hxc",  "mongodb" ] } }

把 { "key" : 26, "values" : [ "mary", "jackson" ] } 整体的返回成一个 count就好了 。。。

{ "_id" : 26, "value" : 2 }
{ "_id" : 29, "value" : 2 }
{ "_id" : 31, "value" : 2 }

——————————————————————————————

 db.t.find()
{ "_id" : 26, "value" : { "key" : 26, "values" : [  "mary",  "jackson" ] } }
{ "_id" : 29, "value" : { "key" : 29, "values" : [  "john",  "redis" ] } }
{ "_id" : 31, "value" : { "key" : 31, "values" : [  "hxc",  "mongodb" ] } }
 db.runCommand({
...      mapreduce:"person",
...      map:function(){
...         emit(this.age,this.name)
...      },
...      reduce:function(key,values){
...         return values.length;
...       },
...       out:"t"
... })
{
	"result" : "t",
	"timeMillis" : 14,
	"counts" : {
		"input" : 6,
		"emit" : 6,
		"reduce" : 3,
		"output" : 3
	},
	"ok" : 1
}
 db.t.find()
{ "_id" : 26, "value" : 2 }    //这个结果就是我们需要的。。。
{ "_id" : 29, "value" : 2 }
{ "_id" : 31, "value" : 2 }

我想把"_id" 替换成 "age", "value" 替换成 “count"

query条件的使用：

> db.t.find()
{ "_id" : 26, "value" : 2 }
{ "_id" : 29, "value" : 2 }
{ "_id" : 31, "value" : 2 }
> db.runCommand({
...      mapreduce:"person",
...      map:function(){
...         emit(this.age,this.name)
...      },
...      reduce:function(key,values){
...         return values.length;
...       },
...       out:"t",
...      finalize:function(key,values){
...        return values;
...      },
...      query:{"$where":function(){return this.age>28}}
... })
{
	"result" : "t",
	"timeMillis" : 69,
	"counts" : {
		"input" : 4,
		"emit" : 4,
		"reduce" : 2,
		"output" : 2
	},
	"ok" : 1
}
> db.t.find()
{ "_id" : 29, "value" : 2 }
{ "_id" : 31, "value" : 2 }