过滤来自多个 collections 的输出

Question

假设我们有这样一组纸条：

 ["xxxxx1","xxxxx2","xxxxx3","xxxxx4"]

我们在数据库中存在以下 collections 个脚本：

Executed:{"_id" : ObjectId("xxx"),"scrip" : "xxxxx1" },{"_id" : ObjectId("xxy"),"scrip" : "xxxxx3" }......
In-process:{"_id" : ObjectId("xxx"),"scrip" : "xxxxx4" }, ....
Rejected:{"_id" : ObjectId("xxx"),"scrip" : "xxxxx5" }....

我们希望获得上述 collections 中不存在的一组凭证。预期输出：

["xxxxx2"]

如何在单个 mongodb pipeline/query 中完成？

Answer 1

以下查询可以获得预期的输出：

db.executed.aggregate([
    {
        $group:{
            "_id":null,
            "executedScrips":{
                $addToSet:"$scrip"
            }
        }
    },
    {
        $lookup:{
            "from":"inprocess",
            "pipeline":[
                {
                    $group:{
                        "_id":null,
                        "inprocessScrips":{
                            $addToSet:"$scrip"
                        }
                    }
                }
            ],
            "as":"inprocessLookup"
        }
    },
    {
        $lookup:{
            "from":"rejected",
            "pipeline":[
                {
                    $group:{
                        "_id":null,
                        "rejectedScrips":{
                            $addToSet:"$scrip"
                        }
                    }
                }
            ],
            "as":"rejectedLookup"
        }
    },
    {
        $unwind:{
            "path":"$inprocessLookup",
            "preserveNullAndEmptyArrays":true
        }
    },
    {
        $unwind:{
            "path":"$rejectedLookup",
            "preserveNullAndEmptyArrays":true
        }
    },
    {
        $project:{  
            "scrips":{
                $concatArrays:[
                    "$executedScrips", 
                    {
                        $ifNull:["$inprocessLookup.inprocessScrips",[]]
                    },
                    {
                        $ifNull:["$rejectedLookup.rejectedScrips",[]]
                    }
                ]
            }
        }
    },
    {
        $project:{
            "_id":0,
            "notFound":{
                $setDifference:[["xxxxx1","xxxxx2","xxxxx3","xxxxx4"],"$scrips"]
            }
        }
    }
]).pretty()

数据集：

Collection: executed

    { "_id" : ObjectId("5d60e572f00e0c8c3593b5ff"), "scrip" : "xxxxx1" }
    { "_id" : ObjectId("5d60e572f00e0c8c3593b600"), "scrip" : "xxxxx3" }

Collection: inprocess

    { "_id" : ObjectId("5d60f23ff00e0c8c3593b601"), "scrip" : "xxxxx4" }

Collection: rejected

    { "_id" : ObjectId("5d60f260f00e0c8c3593b602"), "scrip" : "xxxxx5" }

输出：

{ "notFound" : [ "xxxxx2" ] }

注意：如果executedcollection中没有记录，查询将失败，因为聚合从那里开始。

更新 I：将传递请求数组而不是凭证数组

以下查询可以获得预期的输出：

db.executed.aggregate([
    {
        $group:{
            "_id":null,
            "executedScrips":{
                $addToSet:"$scrip"
            }
        }
    },
    {
        $lookup:{
            "from":"inprocess",
            "pipeline":[
                {
                    $group:{
                        "_id":null,
                        "inprocessScrips":{
                            $addToSet:"$scrip"
                        }
                    }
                }
            ],
            "as":"inprocessLookup"
        }
    },
    {
        $lookup:{
            "from":"rejected",
            "pipeline":[
                {
                    $group:{
                        "_id":null,
                        "rejectedScrips":{
                            $addToSet:"$scrip"
                        }
                    }
                }
            ],
            "as":"rejectedLookup"
        }
    },
    {
        $unwind:{
            "path":"$inprocessLookup",
            "preserveNullAndEmptyArrays":true
        }
    },
    {
        $unwind:{
            "path":"$rejectedLookup",
            "preserveNullAndEmptyArrays":true
        }
    },
    {
        $project:{  
            "scrips":{
                $concatArrays:[
                    "$executedScrips", 
                    {
                        $ifNull:["$inprocessLookup.inprocessScrips",[]]
                    },
                    {
                        $ifNull:["$rejectedLookup.rejectedScrips",[]]
                    }
                ]
            }
        }
    },
    {
        $addFields:{
            "requests":[
                {
                    "requestid" : "R1", 
                    "stocks" : ["xxxxx1","xxxxx2","xxxxx3","xxxxx4"]
                },
                {
                    "requestid" : "R2", 
                    "stocks" : ["xxxxx1","xxxxx3","xxxxx4"]
                },
                {
                    "requestid" : "R3", 
                    "stocks" : ["xxxxx1","xxxxx3","xxxxx4","xxxxx10"]
                }
            ]
        }
    },
    {
        $project:{
            "_id":0,
            "unmatchedRequests":{
               $map:{
                    "input":"$requests",
                    "as":"request",
                    "in":{
                        $concat:{
                            $cond:[
                                {
                                    $gt:[
                                        {
                                            $size:{
                                                $setDifference:["$$request.stocks","$scrips"]
                                            }
                                        },
                                        0
                                    ]
                                },
                                "$$request.requestid",
                                null
                            ]
                        }
                    }
               }
            }
        }
    },
    {
        $project:{
            "unmatchedRequests":{
                $filter:{
                    "input":"$unmatchedRequests",
                    "as":"unmatchedRequest",
                    "cond":{
                        $ne:["$$unmatchedRequest",null]
                    }
                }
            }
        }
    }
]).pretty()

输出：

{ "unmatchedRequests" : [ "R1", "R3" ] }

注意：在第7个聚合阶段，我们正在注入请求数组。

过滤来自多个 collections 的输出

Filter output from multiple collections

mongodb

mongodb-java

mongodb-query

aggregation-framework

数据集：

输出：