MongoDB学习——高级查询

 丰富的查询

MongoDB支持类似于SQL的>，>=，<，<=，<>，分别用 "$gt", "$gte", "$lt", "$lte", "$ne"表示，以及：$in，$nin，也支持条件的组合：and, or。

甚至还支持正则表达式，你没看错，正则表达式，强大吧。

> db.user.find()
{ "_id" : ObjectId("51f730c604f4bacbc87f67f2"), "name" : "jack", "age" : 53, "address" : { "country" : "CN", "city" : "Macao" }, "fav" : [ "apple", "banana" ], "email" : "yeah@yahoo.com" }
{ "_id" : ObjectId("51f7312b04f4bacbc87f67f3"), "name" : "allen", "age" : 20, "address" : { "country" : "US", "city" : "NewYork" }, "email" : "allen@hotmail.com" }
>
> db.user.find({"age":{$lt:30}})
{ "_id" : ObjectId("51f7312b04f4bacbc87f67f3"), "name" : "allen", "age" : 20, "address" : { "country" : "US", "city" : "NewYork" }, "email" : "allen@hotmail.com" }
>
> db.user.find({"address.country":{$in:["CN","US"]},"age":{$gte:30}})
{ "_id" : ObjectId("51f730c604f4bacbc87f67f2"), "name" : "jack", "age" : 53, "address" : { "country" : "CN", "city" : "Macao" }, "fav" : [ "apple", "banana" ], "email" : "yeah@yahoo.com" }
>
> db.user.find({"email":/.+@yahoo.com/})
{ "_id" : ObjectId("51f730c604f4bacbc87f67f2"), "name" : "jack", "age" : 53, "address" : { "country" : "CN", "city" : "Macao" }, "fav" : [ "apple", "banana" ], "email" : "yeah@yahoo.com" }
>

另外MongoDB也提供了where子句，可以定义一个javascipt的predict函数来查询

> db.user.find({$where:function() {return this.address.city == 'Macao'}})
{ "_id" : ObjectId("51f730c604f4bacbc87f67f2"), "name" : "jack", "age" : 53, "address" : { "country" : "CN", "city" : "Macao" }, "fav" : [ "apple", "banana" ], "email" : "yeah@yahoo.com" }
>

 聚集函数

MongoDB也支持常用的聚集函数：count，distinct，group等

其中group较为复杂，需要指定初始值，和$reduce函数，

 另外group还可以加condition指定过滤条件，finalize指定每组文档$reduce完成后的动作

> db.user.find()
{ "_id" : ObjectId("51f7560004f4bacbc87f67f4"), "name" : "brodie", "age" : 30, "gender" : "male", "address" : { "country" : "CN", "city" : "ShangHai" } }
{ "_id" : ObjectId("51f7562304f4bacbc87f67f5"), "name" : "dany", "age" : 28, "gender" : "male", "address" : { "country" : "CN", "city" : "ShangHai" } }
{ "_id" : ObjectId("51f7563d04f4bacbc87f67f6"), "name" : "davis", "age" : 31, "gender" : "male", "address" : { "country" : "CN", "city" : "WuHan" } }
{ "_id" : ObjectId("51f7566904f4bacbc87f67f7"), "name" : "jameson", "age" : 37, "gender" : "female", "address" : { "country" : "US", "city" : "NewYork" } }
>
> db.user.count({"age":{$gte:30}})
3
> db.user.distinct("address.country")
[ "CN", "US" ]
>
> db.user.group({
...     "key":{"address.country":true},
...     "initial":{"users":[]},
...     "condition":{"age":{$gte:30}},
...     "reduce":function(curdoc, predoc) {
...         predoc.users.push(curdoc.name);
...     },
...     "finalize":function(predoc) {
...         predoc.count = predoc.users.length;
...     }
... })
[
    {
        "address.country" : "CN",
        "users" : [
            "brodie",
            "davis"
        ],
        "count" : 2
    },
    {
        "address.country" : "US",
        "users" : [
            "jameson"
        ],
        "count" : 1
    }
]

 MapReduce

MapReduce是比较复杂的聚集函数，但更加灵活。

map函数：指定的分组函数，其中调用emit分组，第一个参数为分组的键，value为reduce时需要的字段；

reduce函数：分组后的处理函数，参数key为分组的key，values为分组后的集合；

mapReduce的第三个参数指定输出的结合名称；

运行mapReduce后会输出执行的状态信息

下面用mapReduce作出上面group同样的效果

 

> var map = function() {
...    emit(this.address.country, {user:this.name,age:this.age,count:1});
... }
> 
> var reduce = function(key, values) {
...     var result = {users:[],count:0};
...     for (var i = 0; i < values.length; ++i) {
...         if (values[i].age >= 30) {
...             result.users.push(values[i].user);
...             result.count += values[i].count;
...         }
...     }
...     return result;
... }
> 
> db.user.mapReduce(map,reduce, {"out":"user_by_country"})
{
	"result" : "user_by_country",
	"timeMillis" : 62, // time used
	"counts" : {
		"input" : 4, // number of input document
		"emit" : 4, // times of emit called
		"reduce" : 1, // times reduce called,note:a key
that has only a single value, will not trigger the reduce
		"output" : 2
	},
	"ok" : 1,
}
> db.user_by_country.find()
{ "_id" : "CN", "value" : { "users" : [  "brodie",  "davis" ], "count" : 2 } }
{ "_id" : "US", "value" : { "user" : "jameson", "age" : 37, "count" : 1 } }
>