1 视图

1.1 创建视图

基本语法

db.createView(
  "<viewName>",
  "<source>",
  [<pipeline>],
  {
    "collation" : { <collation> }
  }
)

#参数解释
<viewName>        必须，视图名称
<source>          必须，数据源，集合/视图
[<pipeline>]      可选，一组管道
"collation"       可选，排序规则

1.2 删除视图

db.视图名.drop()

1.3 例子

t2、t3表数据

> db.t2.find()
{ "_id" : ObjectId("60adbec5d15f5f92b4805399"), "id" : 1, "name" : "a" }
{ "_id" : ObjectId("60adbec5d15f5f92b480539a"), "id" : 2, "name" : "b" }
{ "_id" : ObjectId("60adbec5d15f5f92b480539b"), "id" : 3, "name" : "c" }

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }
{ "_id" : ObjectId("60b0602aac7d71a1301a1715"), "id" : 2, "age" : 20, "city" : "jilin" }

建立视图，基于t2表的id字段，查询t3表id字段相同的内容，t3表查出的内容放到age&city字段中。对于id=2的数据，t3中匹配多行，多行数据放到[ ]中，分隔为2个{ }

db.createView(
  "t2joint3",
  "t2",
  [
    {
      $lookup:
      {
        from:"t3",
        localField:"id",
        foreignField:"id",
        as:"age&city"
      }
    },
    {
      $project:
      {
        "id":1,"name":1,"age&city.age":1,"age&city.city":1
      }
    }
  ]
)

查询视图内容

> db.t2joint3.find()
{ "_id" : ObjectId("60adbec5d15f5f92b4805399"), "id" : 1, "name" : "a", "age&city" : [ { "age" : 10, "city" : "beijing" } ] }
{ "_id" : ObjectId("60adbec5d15f5f92b480539a"), "id" : 2, "name" : "b", "age&city" : [ { "age" : 11, "city" : "shanghai" }, { "age" : 20, "city" : "jilin" } ] }
{ "_id" : ObjectId("60adbec5d15f5f92b480539b"), "id" : 3, "name" : "c", "age&city" : [ { "age" : 12, "city" : "dalian" } ] }

优化
使用$unwind拆分age&city字段中的列表

db.createView(
  "t2joint3",
  "t2",
  [
    {
      $lookup:
      {
        from:"t3",
        localField:"id",
        foreignField:"id",
        as:"age&city"
      }
    },
    {
      $project:
      {
        "id":1,"name":1,"age&city.age":1,"age&city.city":1
      }
    },
    {
      $unwind:
      {
        path:"$age&city"
      }
    }
  ]
)

查询视图内容

> db.t2joint3.find()
{ "_id" : ObjectId("60adbec5d15f5f92b4805399"), "id" : 1, "name" : "a", "age&city" : { "age" : 10, "city" : "beijing" } }
{ "_id" : ObjectId("60adbec5d15f5f92b480539a"), "id" : 2, "name" : "b", "age&city" : { "age" : 11, "city" : "shanghai" } }
{ "_id" : ObjectId("60adbec5d15f5f92b480539a"), "id" : 2, "name" : "b", "age&city" : { "age" : 20, "city" : "jilin" } }
{ "_id" : ObjectId("60adbec5d15f5f92b480539b"), "id" : 3, "name" : "c", "age&city" : { "age" : 12, "city" : "dalian" } }

2 pipeline管道

pipeline将多个文档中的值组合在一起，并可对分组数据执行各种操作，以返回单个结果，相当于SQL中的 count(*)与group by。
aggregation pipeline主要是用aggregate()方法来实现聚合操作。
pipeline通俗的理解就是一个管道，其中每一个操作就是管道的一个阶段，每次当前操作接受上一个阶段的输出作为输入，并把输出结果作为输入结果给下一个阶段。
语法如下：db.collection.aggregate( [ { $语法 }, … ] )

2.1 $project

$project用来指明返回值包含哪些列，包含的置1，不包含的置0（自定义列默认置0，_id列默认置1，需要手工置0）
原数据

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }

只取age列{$project:{"age":1,"_id":0}}

> db.t3.aggregate([{$project:{"age":1,"_id":0}}])
{ "age" : 10 }
{ "age" : 11 }
{ "age" : 12 }

2.2 $match

通过跟查询语句相比对，来过滤集合，只返回跟查询语句相匹配的行。
原数据

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }

只取age=10的行{$match:{"age":10}}

> db.t3.aggregate([{$match:{"age":10}}])
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }

只取age<11的行{$match:{"age":{$lt:11}}}，其余比较符号$gt、$eq、$ne

>  db.t3.aggregate([{$match:{"age":{$lt:11}}}])
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }

2.3 $limit

限制返回的数据条目数
原数据

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }

只取2行{$limit:2}

> db.t3.aggregate([{$limit:2}])
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }

2.4 $skip

跳过n行数据
原数据

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }

跳过2行{$skip:2}

> db.t3.aggregate([{$skip:2}])
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }

2.5 $sort

对所有的输入数据进行排序，并输出排序好的数据。
使用语法

{ $sort: { <field1>: <sort order>, <field2>: <sort order> ... } }

其中sort order可以取值为1，-1：
当sort order取值为1时，代表升序。
当sort order取值为-1时，代表降序。

原数据

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }

按照age列倒序排序{$sort:{"age":-1}}

> db.t3.aggregate([{$sort:{"age":-1}}])
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }

2.6 $unwind

将数据行中含有列表内容的部分，拆分成一个一个数据显示
使用语法

{
  $unwind:
    {
      path: <field path>,
      includeArrayIndex: <string>,
      preserveNullAndEmptyArrays: <boolean>
    }
}


path:必须有,指明需要拆分的字段，一般为一个数组
includeArrayIndex: 可选,如果需要拆分开的数据包含数据下标，可以指定数组下标的字段名
preserveNullAndEmptyArrays: 可选，在数组为null、缺失、空的情况下，如果为true, $unwind 同样输出当前文档. 如果为false, $unwind不输出文档。默认是false.

原数据

> db.t4.find()
{ "_id" : ObjectId("60ae0de2bff5299b604005c6"), "id" : 1, "age" : 10, "list" : [ "a", "b", "c" ] }
{ "_id" : ObjectId("60ae0de2bff5299b604005c7"), "id" : 2, "age" : 11, "list" : [ ] }
{ "_id" : ObjectId("60ae0de2bff5299b604005c8"), "id" : 3, "age" : 12, "list" : null }
{ "_id" : ObjectId("60ae0ec0bff5299b604005c9"), "id" : 4, "age" : 13 }

对list列的数组进行拆分显示，并显示数组下标，下标字段名称为arrayIndex{ $unwind: { path: "$list", includeArrayIndex: "arrayIndex", } }
除第一行外，其余行list字段为空、null、缺失，所以没有显示

> db.t4.aggregate([{   $unwind:     {       path: "$list",       includeArrayIndex: "arrayIndex"    } } ])
{ "_id" : ObjectId("60ae0de2bff5299b604005c6"), "id" : 1, "age" : 10, "list" : "a", "arrayIndex" : NumberLong(0) }
{ "_id" : ObjectId("60ae0de2bff5299b604005c6"), "id" : 1, "age" : 10, "list" : "b", "arrayIndex" : NumberLong(1) }
{ "_id" : ObjectId("60ae0de2bff5299b604005c6"), "id" : 1, "age" : 10, "list" : "c", "arrayIndex" : NumberLong(2) }

配置preserveNullAndEmptyArrays: true

> db.t4.aggregate([ {   $unwind:     {       path: "$list",       includeArrayIndex: "arrayIndex",       preserveNullAndEmptyArrays: true     } } ] )
{ "_id" : ObjectId("60ae0de2bff5299b604005c6"), "id" : 1, "age" : 10, "list" : "a", "arrayIndex" : NumberLong(0) }
{ "_id" : ObjectId("60ae0de2bff5299b604005c6"), "id" : 1, "age" : 10, "list" : "b", "arrayIndex" : NumberLong(1) }
{ "_id" : ObjectId("60ae0de2bff5299b604005c6"), "id" : 1, "age" : 10, "list" : "c", "arrayIndex" : NumberLong(2) }
{ "_id" : ObjectId("60ae0de2bff5299b604005c7"), "id" : 2, "age" : 11, "arrayIndex" : null }
{ "_id" : ObjectId("60ae0de2bff5299b604005c8"), "id" : 3, "age" : 12, "list" : null, "arrayIndex" : null }
{ "_id" : ObjectId("60ae0ec0bff5299b604005c9"), "id" : 4, "age" : 13, "arrayIndex" : null }

2.7 $lookup

将用localField字段中的内容和from集合中的foreignField字段进行比较，如果比较相同，就将from集合中的这条记录加入当前文档中，字段名为as的值。
使用语法

{
   $lookup:
     {
       from: <collection to join>,
       localField: <field from the input documents>,
       foreignField: <field from the documents of the "from" collection>,
       as: <output array field>
     }
}

原数据

> db.t2.find()
{ "_id" : ObjectId("60adbec5d15f5f92b4805399"), "id" : 1, "name" : "a" }
{ "_id" : ObjectId("60adbec5d15f5f92b480539a"), "id" : 2, "name" : "b" }
{ "_id" : ObjectId("60adbec5d15f5f92b480539b"), "id" : 3, "name" : "c" }

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }

基于t2表的id字段联合查询t3表的id字段，将结果合并显示，t3的内容放在location字段中，使用location.city、location.age控制只显示t3中的city和age值

> db.t2.aggregate(
  [
    {
    $lookup:
     {
       from: "t3",
       localField: "id",
       foreignField: "id",
       as: "location"
     }
    },
    {
      $project:
      {
        "id":1,
        "name":1,
        "location.city":1,
        "location.age":1
      }
    }
  ]
)
{ "_id" : ObjectId("60adbec5d15f5f92b4805399"), "id" : 1, "name" : "a", "location" : [ { "age" : 10, "city" : "beijing" } ] }
{ "_id" : ObjectId("60adbec5d15f5f92b480539a"), "id" : 2, "name" : "b", "location" : [ { "age" : 11, "city" : "shanghai" } ] }
{ "_id" : ObjectId("60adbec5d15f5f92b480539b"), "id" : 3, "name" : "c", "location" : [ { "age" : 12, "city" : "dalian" } ] }

2.8 $group

类似与mysql中的group by，指定数据根据某个字段分组。在$group中我们是通过_id来指定分组依据。
原数据

> db.t3.find()
{ "_id" : ObjectId("60adbf29d15f5f92b480539c"), "id" : 1, "age" : 10, "city" : "beijing" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539d"), "id" : 2, "age" : 11, "city" : "shanghai" }
{ "_id" : ObjectId("60adbf29d15f5f92b480539e"), "id" : 3, "age" : 12, "city" : "dalian" }
{ "_id" : ObjectId("60b0602aac7d71a1301a1715"), "id" : 2, "age" : 20, "city" : "jilin" }

指定数据根据id来进行分组，然后在每个分组中通过$sum函数来计算age的总数，并将计算结果给了totalage这个字段。

> db.t3.aggregate([
    {
       $group:{_id:"$id",
              totalage:{$sum:"$age"}
              }
    }
    ])
{ "_id" : 1, "totalage" : 10 }
{ "_id" : 2, "totalage" : 31 }
{ "_id" : 3, "totalage" : 12 }