MongoDB的聚合计算平均并将其添加到文档

Question

我有websites其中包含2个文件：

{ 
    "_id" : ObjectId("58503934034b512b419a6eab"), 
    "website" : "https://www.stackoverflow.com", 
    "name" : "Stack Exchange", 
    "keywords" : [ 
     "helping", 
     "C#", 
     "PYTHON" 
    ] 
} 

{ 
    "_id" : ObjectId("58503934034b512b419a6eab"), 
    "website" : "https://www.google.com.com", 
    "name" : "Stack Exchange", 
    "keywords" : [ 
     "search", 
     "engine", 
    ] 
} 

我也有另外一个seo_tracking其中包含：

{ 
    "_id" : ObjectId("587373d6f6325811c8a0b3ad"), 
    "position" : "2", 
    "real_url" : "https://www.stackoverflow.com", 
    "created_at" : ISODate("2017-01-09T11:28:22.104Z"), 
    "keyword" : "helping" 
}, 
{ 
    "_id" : ObjectId("587373d6f6325811c8a0b3ad"), 
    "position" : "4", 
    "real_url" : "https://www.stackoverflow.com", 
    "created_at" : ISODate("2017-01-09T11:28:22.104Z"), 
    "keyword" : "C#" 
} 

等。这周围包含100 +文件

我想要做的是聚合seo_tracking与website在特定的URL（（websites）将匹配www.stackoverflow.com（seo_tracking）），我可以做得很好。不过，我想回每个websites以下：

{ 
    "_id" : ObjectId("587373d6f6325811c8a0b3ad"), 
    "website":"https://www.stackoverflow.com", 
    "avg_position" : "2" 
} 

那么对于谷歌等。即使avg_position是0 ..我曾尝试以下：

db.seo_tracking.aggregate([ 

    { 
     $lookup: 
     { 
      from: "websites", 
      localField: "real_url", 
      foreignField: "website", 
      as: "post_websites" 
     }, 
    }, 
    { 
     "$group": { 
      _id:null, 
      avg_position:{$avg:"$position"} 
      } 
     } 
]) 

然而，这只是生产：

{ 
    "_id" : null, 
    "avg_position" : 2.0 
} 

我需要做的是website和理想还需要ID

任何想法，我要去哪里错了吗？

Answer 1

你可以尝试这样的事情。你需要$unwind从加入收藏访问领域和更改您的分组键从加入收藏使用_id获得平均每个网站：

db.seo_tracking.aggregate([{ 
    $lookup: { 
     from: "website", 
     localField: "real_url", 
     foreignField: "website", 
     as: "post_websites" 
    }, 
}, { 
    $unwind: "$post_websites" 
}, { 
    "$group": { 
     _id: "$post_websites._id", 
     avg_position: { 
      $avg: "$position" 
     }, 
     website: { 
      $first: "$real_url" 
     } 
    } 
}])