按每个父 _id 过滤嵌入的文档？

Question

我有文件

{
    "_id" : ObjectId("5da832caeb173112348e509b"),
    "owner" : {
        "image" : "5d999578aeb073247de4bd6e.jpg",
        "fullname" : "hem sopheap",
        "userID" : "5d999578aeb073247de4bd6e"
    },
    "project" : {},
    "image" : "hem sopheap-1571304138866.png",
    "body" : "Lorem Ipsum "),
    "comments" : [ 

        {
            "user" : "5d999578aeb073247de4bd6e",
            "fullname" : "hem sopheap",
            "username" : "sopheap",
            "comment" : "1000000",
            "_id" : ObjectId("5db07900ae100b0c05b1222c"),
            "replies" : [],
            "date" : ISODate("2019-10-23T15:40:57.535Z"),
            "likes" : [ 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044", 
                "5da85558886aee13e4e7f044"
            ]
        }, 
        {
            "user" : "5d999578aeb073247de4bd6e",
            "fullname" : "hem sopheap",
            "username" : "sopheap",
            "comment" : "11111111111",
            "_id" : ObjectId("5db0790aae100b0c05b1222d"),
            "replies" : [],
            "date" : ISODate("2019-10-23T15:40:57.535Z"),
            "likes" : []
        }
    ],
    "__v" : 33,
    "likes" : [ 
        "5d999578aeb073247de4bd6e"
    ]
}

我怎样才能得到likes，通过帖子_id和cmets_id过滤得到结果likes

"likes" : [ 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044", 
                    "5da85558886aee13e4e7f044"
                ]

Answer 1

让我们简化对此的输入。我们有两个帖子，P0 和 P2。每个都有一个数组comments。我们知道comment._id 至少在帖子中是唯一的，因此可以在这里“重用”它们：

var r =
[
 {
    "_id" : "P0",
    "comments" : [
{ "_id" : "C0", "likes" : [ "AA", "AA" ] }
,{"_id" : "C1", "likes" : [] }
,{"_id" : "C2",  "likes" : [ "foo", "bar" ]}
,{"_id" : "C3",  "likes" : []
        }
    ]
 }

 ,{
    "_id" : "P2",
    "comments" : [
{ "_id" : "C0", "likes" : [ "FF", "FF" ] }
,{"_id" : "C1", "likes" : [] }
,{"_id" : "C2",  "likes" : [ "foo", "bar" ]}
,{"_id" : "C3",  "likes" : []
        }
    ]
 }
];

这里有一个解决方案：

db.foo.aggregate([
// First, match on post ID and comments ID.  Remember, comments is an           
// array so ANY comments entry with key C0 inside the array will match and      
// yield the entire array.  But this is OK because it very much narrows down    
// the info to process:                                                         
{$match: {_id: "P2", "comments._id":"C0"}}

,{$unwind: "$comments"} // Unwind the comments:                                 

// And now pick only that comment with ID C0:                                   
,{$match: {"comments._id":"C0"}}

// To complete the request, make "likes" a top level field:                     
,{$project: {"likes": "$comments.likes"}}

   ]);

对 OP 的完整答案将包括将帖子和评论 _id 设置为 ObjectId 而不是字符串，但查询是相同的。