如何从mongodb分组连接的文档（图的节点）

Question

Assume that I have a one-way graph with nodes A,B,C,D,E such that

A->B->C is a connected component,

D->E is another connected component.

Nodes are saved in MongoDB as documents

{name:'A', child: 'B'}, {name:'B', child: 'C'}, {name:'C'},
{name:'D', child: 'E'}, {name:'E'}

How to get all connected components?

Expected result: 2 groups

[{name:'A'...},{name:'B'...}, {name:'C'...}],[{name:'D'...}, {name:'E'...}]

Answer 1

Use $graphLookup pipeline stage.

• $group - collects all possible values from the name and child fields.
• $addFields - generates the array with name field values of the root documents.
• $unwind - splits roots array into seperated documents.
• $graphLookup - collects related documents by given name field value of the root document.
• $project - removes the not necessary fieldss from the result documents.

Query:

db.getCollection('t').aggregate([
    {
        $group: {
            _id: null,
            names: { $addToSet: "$name" },
            childs: { $addToSet: "$child" }
        }
    },
    {
        $addFields: {
            roots: { $setDifference: ["$names", "$childs"] }
        }
    },
    { $unwind: "$roots" },
    {
        $graphLookup: {
            from: "t",
            startWith: "$roots",
            connectFromField: "child",
            connectToField: "name",
            as: "related"
        }
    },
    {
        $project: {
            "related": 1,
            "_id": 0
        }
    }
])

Result:

/* 1 */
{
    "related" : [ 
        {
            "_id" : ObjectId("5a29316a545eb40950c33bc8"),
            "name" : "C"
        }, 
        {
            "_id" : ObjectId("5a29316a545eb40950c33bc7"),
            "name" : "B",
            "child" : "C"
        }, 
        {
            "_id" : ObjectId("5a29316a545eb40950c33bc6"),
            "name" : "A",
            "child" : "B"
        }
    ]
}

/* 2 */
{
    "related" : [ 
        {
            "_id" : ObjectId("5a29316a545eb40950c33bca"),
            "name" : "E"
        }, 
        {
            "_id" : ObjectId("5a29316a545eb40950c33bc9"),
            "name" : "D",
            "child" : "E"
        }
    ]
}