MongoDB聚合

MongoDB
数据库
发布日期:   2018-12-11
文章字数:   1.4k
阅读时长:   6 分
阅读次数:  

distinct

distinct用来找出给定键的所有不同的值。使用时必须指定集合和键。

元数据
> db.user.find()
{ "_id" : ObjectId("5c0df99fbc6d47cbcdb55fd0"), "name" : "jack", "age" : 19 }
{ "_id" : ObjectId("5c0df9abbc6d47cbcdb55fd1"), "name" : "rose", "age" : 20 }
{ "_id" : ObjectId("5c0df9b0bc6d47cbcdb55fd2"), "name" : "jack", "age" : 18 }
{ "_id" : ObjectId("5c0df9c3bc6d47cbcdb55fd3"), "name" : "tony", "age" : 21 }
{ "_id" : ObjectId("5c0df9cdbc6d47cbcdb55fd4"), "name" : "adam", "age" : 18 }
{ "_id" : ObjectId("5c0e0824bc6d47cbcdb55fd5"), "age" : 2, "name" : 1 }
{ "_id" : ObjectId("5c0e0826bc6d47cbcdb55fd6"), "age" : 2, "name" : 2 }
{ "_id" : ObjectId("5c0e0828bc6d47cbcdb55fd7"), "age" : 2, "name" : 3 }
{ "_id" : ObjectId("5c0f3476478f8e67a82bc840"), "name" : "jack", "age" : 19 }

> db.runCommand({distinct:"user","key":"age"})
{ "values" : [ 2, 18, 19, 20, 21 ], "ok" : 1 }

group

group做的聚合稍复杂一些。先选定分组所依据的键，而后MongoDB就会将集合依据选定键值的不同分成若干组。然后可以通过聚合每一组内的文档，产生一个结果文档。

元数据
{"day":"2010/10/03","time":"10/3/2010 03:57:01 GMT-400","price":4.23}
{"day":"2010/10/04","time":"10/4/2010 11:45:01 GMT-400","price":4.27}
{"day":"2010/10/05","time":"10/5/2010 05:43:01 GMT-400","price":4.11}
{"day":"2010/10/06","time":"10/6/2010 06:56:01 GMT-400","price":4.01}

group查询语句

> db.runCommand({
...     "group": {
...         "ns": "stocks",
...         "key": "day",
...         "inital": {
...             "time": 0
...         },
...         "$reduce": function(doc, prev) {
...             if (doc.time > prev.time) {
...                 prev.price = doc.price;
...                 prev.time = doc.time;
...             }
...         }
...     }
... })
{
        "ok" : 0,
        "errmsg" : "initial has to be an object",
        "code" : 2,
        "codeName" : "BadValue"
}

“ns”: “stocks” 指定进行分组的集合

“key”: “day”, 指定文档分组一句的键,这里就是”day”,所有”day”值相同的w文档被划分到了一组,

“inital”: { “time”: 0 } 每一组reduce函数调用的初始时间,会作为初始文档,传递给后续过程,每一个组员的所有成员都是用这个累加器,所以改变会保留住.

“$reduce”: function(doc, prev)每个文档都对应一次这个调用。系统会传递两个参数：当前文档和累加器文档(本组当前的结果）。本例中，想让reduce函数比较当前文档的时间和累加器的时间。如果当前文档的时间更近，则将累加器的日期和价格替换成当前文档的值。别忘了，每一组都有一个独立的累加器，所以不必担心不同的日斯使用同一个累加器。

如果只要最近30天的股价可以使添加condition

> db.runCommand({
...     "group": {
...         "ns": "stocks",
...         "key": "day",
...         "inital": {
...             "time": 0
...         },
...         "$reduce": function(doc, prev) {
...             if (doc.time > prev.time) {
...                 prev.price = doc.price;
...                 prev.time = doc.time;
...             }
...         },
... "condition":{"day":{$gt:"2010/09/30"}}
...
...     }
... })

这里每组的”price”都是显式设置的，”time”先由初始化器设置，然后也是主动更新。”day”是默认被加进去的，因为分组依据的键默认被加入到每个”retval”内嵌文档中。要是不想返回这个键，可以用完成器把累加器文档变成任意形态，甚
至变换成非文档（例如数字或字符串）。

aggregate

aggregate 提供的是类似SQL（结构化查询语言）的聚合操作，例如每个操作符都可以找到对应的sql关键字

MySQL MongoDB
WHERE $match
GROUP BY $group
HAVING $match
SELECT $project
ORDERY BY $sort
LIMIT $limit
SUM $sum
COUNT() $sum $SortByCount
join $lookup

sql语句 与对应的聚合函数
FYGv1U.md.png

MapReduce

FYJJgS.md.png

Python 脚本导入数据

from pymongo import MongoClient
from random import randint
import datetime

client = MongoClient()
db = client.get_database('taobao')

order = db.order_info

status = ['A', 'B', 'C']

cust_id = ['A123', 'B123', 'C123']

price = [500, 250, 200, 300]

sku = ['mmm', 'nnn']

for i in range(1, 100):
    items = []
    items_count = randint(2, 6)
    for n in range(items_count):
        # sku 库存量  qty 数量
        items.append({"sku": sku[randint(0, 1)], "qty": randint(1, 10), "price": randint(0, 5)})
    new = {
        "status": status[randint(0, 2)],
        "cust_id": cust_id[randint(0, 2)],
        "price": price[randint(0, 3)],
        "ord_date": datetime.datetime.utcnow(),
        "items": items
    }
    print(new)
    order.insert_one(new)
    print(i)
print(order.estimated_document_count())

数据格式

{
        "_id" : ObjectId("5c0f1bc52a3cde1260163371"),
        "status" : "B",
        "cust_id" : "C123",
        "price" : 300,
        "ord_date" : ISODate("2018-12-11T02:07:01.598Z"),
        "items" : [
                {
                        "sku" : "nnn",
                        "qty" : 2,
                        "price" : 5
                },
                {
                        "sku" : "mmm",
                        "qty" : 1,
                        "price" : 4
                }
        ]
}

查询每个cust_id 的所有price总和MapReduce

> #定义 map函数
> var mapFunction1 = function() {
...                        emit(this.cust_id, this.price);
...                    };

> #定义reduce函数
> var reduceFunction1 = function(keyCustId, valuesPrices) {
...                           return Array.sum(valuesPrices);
...                       };

> #执行mapreduce，输出到当前db的map_reduce_example集合中
> db.order_info.mapReduce(
...                      mapFunction1,
...                      reduceFunction1,
...                      { out: "map_reduce_example" }
...                    )
{
        "result" : "map_reduce_example",
        "timeMillis" : 284,
        "counts" : {
                "input" : 99,
                "emit" : 99,
                "reduce" : 3,
                "output" : 3
        },
        "ok" : 1
}

查看结果

> db.map_reduce_example.find()
{ "_id" : "A123", "value" : 8350 }
{ "_id" : "B123", "value" : 9150 }
{ "_id" : "C123", "value" : 12800 }
>

聚合管道操作命令

> db.order_info.aggregate({ $group: { _id: "$cust_id", total: { $sum: "$price" }}})
{ "_id" : "A123", "total" : 8350 }
{ "_id" : "B123", "total" : 9150 }
{ "_id" : "C123", "total" : 12800 }
>

计算所有items 的平均库存 Mapreduce

# map函数
> var mapFunction2 = function() {
...                        for (var idx = 0; idx < this.items.length; idx++) {
...                            var key = this.items[idx].sku;
...                            var value = {
...                                          count: 1,
...                                          qty: this.items[idx].qty
...                                        };
...                            emit(key, value);
...                        }
...                     };
>

    
#reduce函数

> var reduceFunction2 = function(keySKU, countObjVals) {
...                      reducedVal = { count: 0, qty: 0 };
...
...                      for (var idx = 0; idx < countObjVals.length; idx++) {
...                          reducedVal.count += countObjVals[idx].count;
...                          reducedVal.qty += countObjVals[idx].qty;
...                      }
...
...                      return reducedVal;
...                   };

#finalize函数

> var finalizeFunction2 = function (key, reducedVal) {
...
...                        reducedVal.avg = reducedVal.qty/reducedVal.count;
...
...                        return reducedVal;
...
...                     };
>
# 执行mapreduce
> db.order_info.mapReduce( mapFunction2,
...                      reduceFunction2,
...                      {
...                        out: { merge: "map_reduce_example" },
...                        finalize: finalizeFunction2
...                      }
...                    )
{
        "result" : "map_reduce_example",
        "timeMillis" : 121,
        "counts" : {
                "input" : 99,
                "emit" : 406,
                "reduce" : 2,
                "output" : 5
        },
        "ok" : 1
}

查看

> db.map_reduce_example.find()
{ "_id" : "A123", "value" : 8350 }
{ "_id" : "B123", "value" : 9150 }
{ "_id" : "C123", "value" : 12800 }
{ "_id" : "mmm", "value" : { "count" : 211, "qty" : 1135, "avg" : 5.37914691943128 } }
{ "_id" : "nnn", "value" : { "count" : 195, "qty" : 1016, "avg" : 5.21025641025641 } }

聚合管道操作命令实现，计算其所有items 的平均库存,要求输出结果包含count和qty；

> db.order_info.aggregate({$unwind:"$items"},{$group:{_id:"$items.sku",count:{$sum:1},totallyqty:{"$sum":"$items.qty"},avgsku:{"$avg":"$items.qty"}}})
{ "_id" : "nnn", "count" : 195, "totallyqty" : 1016, "avgsku" : 5.21025641025641 }
{ "_id" : "mmm", "count" : 211, "totallyqty" : 1135, "avgsku" : 5.37914691943128 }

用聚合管道操作命令实现：根据cust_id，仓库编号进行分组，计算其所有items 的平均库存；

> db.order_info.aggregate({$unwind:"$items"},{$group:{_id:{cust_id:'$cust_id',skunn:'$items.sku'},avgsku:{"$avg":"$items.qty"}}})
{ "_id" : { "cust_id" : "B123", "skunn" : "mmm" }, "avgsku" : 5.283783783783784 }
{ "_id" : { "cust_id" : "B123", "skunn" : "nnn" }, "avgsku" : 5.121212121212121 }
{ "_id" : { "cust_id" : "C123", "skunn" : "nnn" }, "avgsku" : 5.216216216216216 }
{ "_id" : { "cust_id" : "A123", "skunn" : "nnn" }, "avgsku" : 5.3090909090909095 }
{ "_id" : { "cust_id" : "A123", "skunn" : "mmm" }, "avgsku" : 5.508196721311475 }
{ "_id" : { "cust_id" : "C123", "skunn" : "mmm" }, "avgsku" : 5.368421052631579 }
文章作者: 清风笑丶
文章链接: http://www.hphblog.cn/2018/12/11/mongodb-ju-he/
版权声明: 本博客所有文章除特別声明外，均采用 CC BY 4.0 许可协议。转载请注明来源 清风笑丶 !
MongoDB
 上一篇
Elasticsearch简介与安装 Elasticsearch简介与安装
搜索就是在任何场景下，找寻你想要的信息，这个时候，会输入一段你要搜索的关键字，然后就期望找到这个关键字相关的有些信息 垂直搜索站内搜索 互联网搜索电商网站，招聘网站，新闻网站，各种app IT系统的搜索OA软件，办公自动化软件，会议管理，日
2018-12-15 NOSQL
Elasticsearch
下一篇 
MongoDB索引 MongoDB索引
简介索引就是用来加速查询的。数据库索引与书籍的索引类似：有了索引就不需要翻遍整本书，数据库则可以直接在索引中査找，使得查找速度能提高几个数量级。在索引中找到条目以后，就可以直接跳转到目标文档的位置。让这个比喻走个极端，可以说创建数据库索引就
2018-12-10 数据库
MongoDB
  目录