如何在 mongoDB 中取消嵌套和分组集合
Posted
技术标签:
【中文标题】如何在 mongoDB 中取消嵌套和分组集合【英文标题】：How to un-nest and group collections in mongoDB 【发布时间】：2021-06-18 12:32:54 【问题描述】：我不明白如何在 mongoDB 中展开然后嵌套集合。基本上我有两个结构如下的集合：
问题文档：
"_id" : 1,
"questions" : [
"_id" : 1,
"body" : "What fabric is the top made of?",
"date_written" : "2018-01-04",
"asker_name" : "yankeelover",
"asker_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 2
,
"_id" : 2,
"body" : "HEY THIS IS A WEIRD QUESTION!!!!?",
"date_written" : "2019-04-28",
"asker_name" : "jbilas",
"asker_email" : "first.last@gmail.com",
"reported" : 1,
"helpful" : 4
,
"_id" : 4,
"body" : "How long does it last?",
"date_written" : "2019-07-06",
"asker_name" : "funnygirl",
"asker_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 6
,
答案文档：
"_id" : 1,
"answers" : [
"_id" : 8,
"body" : "DONT BUY IT! It's bad for the environment",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 8
,
"_id" : 7,
"body" : "Its the best! Seriously magic fabric",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 7
,
"_id" : 5,
"body" : "Something pretty soft but I can't be sure",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 5,
"photos" : [
"_id" : 1,
"url" : "https://images.unsplash.com/photo-1530519729491-aea5b51d1ee1?ixlib=rb-1.2.1&ixid=eyJhcHBfaWQiOjEyMDd9&auto=format&fit=crop&w=1651&q=80"
,
答案文档中的 _id 字段与作为答案的问题的 _id 字段相匹配。
最终目标是获得如下所示的数据：
"_id": "17762",
"questions": [
"question_id": 152829,
"question_body": "Why Does it look like this?",
"question_date": "2021-03-06T00:00:00.000Z",
"asker_name": "garethTheGreato",
"question_helpfulness": 60,
"reported": false,
"answers":
"1443770":
"id": 1443770,
"body": "This question was really helpful! Thank you.",
"date": "2021-03-09T00:00:00.000Z",
"answerer_name": "SatisfiedCustomer",
"helpfulness": 3,
"photos": []
,
"1443807":
"id": 1443807,
"body": "mimk",
"date": "2021-03-09T00:00:00.000Z",
"answerer_name": "jij",
"helpfulness": 3,
"photos": [
"blob:http://localhost:3000/8f6375b3-0795-4210-bef7-f112feed8244"
]
,
"1443834":
"id": 1443834,
"body": "10/10 would recomend.",
"date": "2021-03-09T00:00:00.000Z",
"answerer_name": "Krista",
"helpfulness": 2,
"photos": []
,
"1443845":
"id": 1443845,
"body": "Thank you so much for playing my game!",
"date": "2021-03-10T00:00:00.000Z",
"answerer_name": "itsameemario",
"helpfulness": 1,
"photos": []
,
"1443880":
"id": 1443880,
"body": "Tree",
"date": "2021-03-10T00:00:00.000Z",
"answerer_name": "Tree",
"helpfulness": 0,
"photos": [
"blob:http://localhost:3000/123051b6-4dfb-410a-a96f-d4a5128e3056"
]
,
"question_id": 152702,
"question_body": "Please write your question here",
"question_date": "2021-03-05T00:00:00.000Z",
"asker_name": "Your nickname",
"question_helpfulness": 32,
"reported": false,
"answers":
,
我遇到的问题是，当我运行查找时，我得到了一个与问题集合相关的答案数组，但由于答案深深嵌套，我不确定如何获得他们特定问题的每组答案.
这是我目前所拥有的：（暂时忽略切片和排序，这些是我稍后作为项目的另一部分需要用到的参数）
db.prodquests.aggregate([
$match: _id: 5 ,
$unwind: '$questions' ,
$match: 'questions.reported': $lt: 1 ,
$sort: 'questions.helpful': -1 ,
$group: _id: '$_id', questions: $push: '$questions' ,
$project: _id: 1, questions: $slice: ['$questions', 0, 1] ,
$unwind: '$questions' ,
$lookup:
from: 'groupansphotos',
localField: 'questions._id',
foreignField: '_id',
as: 'answers',
,
,
])
这条语句的返回如下：
"_id" : 5,
"questions" :
"_id" : 37,
"body" : "Why is this product cheaper here than other sites?",
"date_written" : "2018-10-18",
"asker_name" : "willsmith",
"asker_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 4
,
"answers" : [
"_id" : 37,
"answers" : [
"_id" : 68,
"body" : "We are selling it here without any markup from the middleman!",
"date_written" : "2018-08-18",
"answerer_name" : "Seller",
"answerer_email" : "null",
"reported" : 0,
"helpful" : 4
]
]
本质上，我只想将答案数组分组到 _id 字段匹配的相应问题下。
提前谢谢你！
【问题讨论】：
【参考方案1】：更新基于 cmets：
更新查询：
db.questions.aggregate([
$match: _id: 5 ,
$unwind: '$questions' ,
$match: 'questions.reported': $lt: 1 ,
$sort: 'questions.helpful': -1 ,
$lookup:
from: "answers",
let: question_id: "$questions._id" ,
pipeline: [
$match:
$expr: $eq: ["$_id", "$$question_id"]
,
$unwind: "$answers" ,
$project:
_id: 0,
k: $toString: "$answers._id" ,
v: "$$ROOT.answers"
],
as: "answers"
,
$group:
_id: "$_id",
questions:
$push:
question_id: "$questions._id",
question_body: "$questions.body",
question_date: "$questions.date_written",
asker_name: "$questions.asker_name",
question_helpfulness: "$questions.helpful",
reported: "$questions.reported",
answers: $arrayToObject: "$answers"
]);
旧查询：
注意：请修正集合名称和/或字段名称。试试这个查询：
db.questions.aggregate([
$match: _id: 5 ,
$unwind: '$questions' ,
$match: 'questions.reported': $lt: 1 ,
$sort: 'questions.helpful': -1 ,
$lookup:
from: "answers",
let: question_id: "$questions._id" ,
pipeline: [
$match:
$expr: $eq: ["$_id", "$$question_id"]
,
$unwind: "$answers" ,
$project:
_id: 0,
k: $toString: "$answers._id" ,
v: "$$ROOT.answers"
],
as: "answers"
,
$match:
$expr:
$gt: [ $size: "$answers" , 0]
,
$group:
_id: "$_id",
questions:
$push:
question_id: "$questions._id",
question_body: "$questions.body",
question_date: "$questions.date_written",
asker_name: "$questions.asker_name",
question_helpfulness: "$questions.helpful",
reported: "$questions.reported",
answers: $arrayToObject: "$answers"
]);
输出：
"_id" : 5,
"questions" : [
"question_id" : 2,
"question_body" : "HEY THIS IS A WEIRD QUESTION!!!!?",
"question_date" : "2019-04-28",
"asker_name" : "jbilas",
"question_helpfulness" : 4,
"reported" : 0,
"answers" :
"14" :
"_id" : 14,
"body" : "DONT BUY IT! It's bad for the environment",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 8
,
"15" :
"_id" : 15,
"body" : "Its the best! Seriously magic fabric",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 7
,
"16" :
"_id" : 16,
"body" : "Something pretty soft but I can't be sure",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 5
,
"question_id" : 1,
"question_body" : "What fabric is the top made of?",
"question_date" : "2018-01-04",
"asker_name" : "yankeelover",
"question_helpfulness" : 2,
"reported" : 0,
"answers" :
"11" :
"_id" : 11,
"body" : "DONT BUY IT! It's bad for the environment",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 8
,
"12" :
"_id" : 12,
"body" : "Its the best! Seriously magic fabric",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 7
,
"13" :
"_id" : 13,
"body" : "Something pretty soft but I can't be sure",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 5
]
测试数据：
questions收藏
"_id" : 5,
"questions" : [
"_id" : 1,
"body" : "What fabric is the top made of?",
"date_written" : "2018-01-04",
"asker_name" : "yankeelover",
"asker_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 2
,
"_id" : 2,
"body" : "HEY THIS IS A WEIRD QUESTION!!!!?",
"date_written" : "2019-04-28",
"asker_name" : "jbilas",
"asker_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 4
,
"_id" : 4,
"body" : "How long does it last?",
"date_written" : "2019-07-06",
"asker_name" : "funnygirl",
"asker_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 6
]
answers收藏：
/* 1 */
"_id" : 1,
"answers" : [
"_id" : 11,
"body" : "DONT BUY IT! It's bad for the environment",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 8
,
"_id" : 12,
"body" : "Its the best! Seriously magic fabric",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 7
,
"_id" : 13,
"body" : "Something pretty soft but I can't be sure",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 5
]
,
/* 2 */
"_id" : 2,
"answers" : [
"_id" : 14,
"body" : "DONT BUY IT! It's bad for the environment",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 8
,
"_id" : 15,
"body" : "Its the best! Seriously magic fabric",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 7
,
"_id" : 16,
"body" : "Something pretty soft but I can't be sure",
"date_written" : "2018-01-04",
"answerer_name" : "metslover",
"answerer_email" : "first.last@gmail.com",
"reported" : 0,
"helpful" : 5
]
【讨论】：
这正是我想要的，非常感谢！在查找阶段拥有管道似乎非常强大，我希望像你一样掌握它。 我们不能从$lookup 阶段内部执行此操作。从外部，我们可以在连接操作产生的字段上使用$unwind，以从外部/本地集合中排除文档，但在您的情况下，我们不能使用$unwind，因为我们需要该数组将其转换为对象。
对不起，我应该更清楚。现在，当我对所选问题执行联接时，如果问题没有匹配的答案文档，则将从最终结果中删除。我想包括在查找之前从原始匹配阶段返回的所有问题，如果它们存在的话，则与他们的答案一起加入
我已经更新了答案。检查它是否有帮助。刚刚删除了$group 之前的$match 阶段。
哇，这是一个非常简单的修复。我现在可以看到那肯定是看错了问题。非常感谢您的帮助！！！以上是关于如何在 mongoDB 中取消嵌套和分组集合的主要内容，如果未能解决你的问题，请参考以下文章
如何根据从 objectId 中提取的创建日期对 MongoDB 集合进行分组？