如何比较同一集合中的文档？

Question

我是 mongo 的新手，因为我以前只使用 Oracle 数据库。我有一个 mongo 数据库，它在列中存储 bitbucket 数据，如下所示：

_id | _class | collectorItemId| firstEverCommit | scmUrl | scmBranch | scmAuthor | scmCommitTimestamp

为了节省时间，我省略了其中的一些专栏。对于 scmBranch 列，该列填充有两个字符串之一："master" 或 "develop"。以下是数据的示例：

这是其中一行的文档视图：

{
"_id" : ObjectId("5e39d6a0330c130006a042c6"),
"collectorItemId" : ObjectId("5e33a6b9887ef5000620a0c0"),
"firstEverCommit" : false,
"scmUrl" : "sampleRepo1",
"scmBranch" : "master",
"scmRevisionNumber" : "a2ad6842468eb55bffcbe7d700b6addd3eb11629",
"scmAuthor" : "son123",
"scmCommitTimestamp" : NumberLong(1580841662000)
}

我现在正在尝试制定 mongo 查询以获得以下数据：

 1. For each scmUrl, If max(scmCommitTimestamp) where scmBranch =
    "develop" > max(scmCommitTimestamp) where scmBranch = "master" THEN
    count the number of rows (i.e commits) where scmBranch = "develop"
    AND scmCommitTimestamp > max(scmCommitTimestamp) where scmBranch =
    "master"

 2. For the results found in #1, find the oldest commit and newest
    commit

到目前为止，我能想到的最好的 mongo 查询如下：

db.bitbucket.aggregate([{
    "$group": {
        "_id": {
            "scmUrl": "$scmUrl",
            "scmBranch": "$scmBranch"
        },
        "MostRecentCommit": {
            "$max": {"$toDate":"$scmCommitTimestamp"}
        }
    }
},{
    "$project": {
        "RepoName": {"$substr": ["$_id.scmUrl",39,-1]},
        "Branch": "$_id.scmBranch",
        "MostRecentCommit": "$MostRecentCommit"
    }
},{
   "$sort":{
       "RepoName":1,
       "Branch":1
       }

}
])

但这只会让我返回每个 scmUrl（即回购）的 develop 分支和 master 分支的最新提交，如下所示：

理想情况下，我想返回 table 个包含以下列的结果：

scmUrl/RepoName | Number of commits on develop branch that are not on master branch| oldest commit in develop branch that's not in master branch | newest commit in develop branch that's not in master branch

如何修改我的 mongo 查询以提取我想要的数据？

Answer 1

你可以尝试这样的事情。

下面的查询将从 master 获取每个 repo 的最新提交日期。在您获得最新的提交日期之后，您将重新加入同一个集合以拉取分支正在开发的所有提交，并且每个 repo 的提交都比主分支更新。

db.bitbucket.aggregate([
  {"$match":{"scmBranch":"master"}},
  {"$group":{"_id":"$scmUrl","recentcommitdate":{"$max":"$scmCommitTimestamp"}}},
  {"$lookup":{
   "from":"bitbucket",
    "let":{"scmUrl":"$_id","recentcommitdate":"$recentcommitdate"},
    "pipeline":[
      {"$match":{"$expr":
        {"$and":[
          {"$eq":["$scmBranch","develop"]},
          {"$eq":["$scmUrl","$$scmUrl"]},
          {"$gte":["$scmCommitTimestamp", "$$recentcommitdate"]}
        ]}
      }},
      {"$sort":{"scmCommitTimestamp":-1}}
    ],
  "as":"commits"
  }},
  {"$match":{"commits.0":{"$exists":true}}},
  {"$project":{
     "commits":{"$size":"$commits"},
     "lastcommit":{"$arrayElemAt":["$commits",0]},
     "firstcommit":{"$arrayElemAt":["$commits",-1]}
  }}
])

此处添加示例https://mongoplayground.net/p/wLnFY0H_nJz

更新版本 ID

db.bitbucket.aggregate([
  {"$match":{"scmBranch":"master"}},
  {"$group":{"_id":"$scmUrl","revisionIds":{"$push":"$scmRevisionNumber"}}},
  {"$lookup":{
   "from":"bitbucket",
    "let":{"scmUrl":"$_id","revisionIds":"$revisionIds"},
    "pipeline":[
      {"$match":{"$expr":
        {"$and":[
          {"$eq":["$scmBranch","develop"]},
          {"$eq":["$scmUrl","$$scmUrl"]},
          {"$not":[{"$in":["$scmRevisionNumber","$$revisionIds"]}]}
        ]}
      }},
      {"$sort":{"scmCommitTimestamp":-1}}
    ],
  "as":"commits"
  }},
  {"$match":{"commits.0":{"$exists":true}}},
  {"$project":{
     "commits":{"$size":"$commits"},
     "lastcommit":{"$arrayElemAt":["$commits",0]},
     "firstcommit":{"$arrayElemAt":["$commits",-1]}
  }}
])

如何比较同一集合中的文档？

How to compare documents in the same collection?

mongodb

nosql

mongodb-query

nosql-aggregation

aggregation-framework