嵌套对象上的 MongoDB java 驱动程序聚合答案

【问题标题】：MongoDB java driver aggregation over nested objects嵌套对象上的 MongoDB java 驱动程序聚合
【发布时间】：2015-03-08 19:32:47
【问题描述】：

我在 mongodb 中有一个集合，其条目格式为：

    {  
   "_id":ObjectId("54fb1dec44ae41b348e22545"),
   "dateLog":   ISODate("2015-03-07T16:00:00   Z"),
   "totalHours":0,
   "projects":[  
      {  
         "client":"BB",
         "name":"Training",
         "component":"ElasticSearch",
         "region":"test",
         "hours":1
      },
      {  
         "client":"PAC",
         "name":"TSO",
         "component":"feature1",
         "region":"a",
         "hours":2
      },
      {  
         "client":"PAC",
         "name":"TSO",
         "component":"feature1",
         "region":"b",
         "hours":2
      }
   ],
   "userName":"guest"
}

我想执行一个聚合查询并计算上面所有条目的 projects.client、projects.name 和 projects.component 的数量。到目前为止，我的聚合查询类似于：

    db.worklog.aggregate( 
{  
   $match:{  
      userName:"guest"
   }
},
{  
   $project:{  
      "projects.client":1,
      "projects.name":1,
      "projects.component":1
   }
},
{  
   $group:{  
      _id:"$projects.client",
      sum:{  
         $sum:"$projects.client"
      }
   }
}
 )

但结果是{ "_id" : [ "BB", "PAC", "PAC" ], "sum" : 0 }

我想算的是：

{ "_id" : { "projects.component" : "ElasticSearch", "projects.client" : "BB" , "projects.name" : "Training"} , "count" : 1}
{ "_id" : { "projects.component" : "feature1", "projects.client" : "PAC" , "projects.name" : "TSO"} , "count" : 1}

正确的查询是什么？

【问题讨论】：

标签： mongodb

【解决方案1】：

通过使用下面的sn-p和mongo java驱动，我想我可以得到想要的结果：

DBObject match = new BasicDBObject("$match", new BasicDBObject("userName", userName));

    DBObject fields = new BasicDBObject("compkey", 1);
    fields.put("projects.client", 1);
    fields.put("projects.component", 1);
    fields.put("projects.name", 1);
    fields.put("_id", 0);
    DBObject project = new BasicDBObject("$project", fields );

    Map<String, Object> dbObjIdMap = new HashMap<String, Object>();
    dbObjIdMap.put("projects.client", "$projects.client");
    dbObjIdMap.put("projects.component", "$projects.component");
    dbObjIdMap.put("projects.name", "$projects.name");

    DBObject groupFields = new BasicDBObject( "_id", new BasicDBObject(dbObjIdMap));
    groupFields.put("count", new BasicDBObject( "$sum", 1));
    DBObject group = new BasicDBObject("$group", groupFields);
    DBObject sort = new BasicDBObject("$sort", new BasicDBObject("count", -1));
    DBObject limit = new BasicDBObject("$limit", count);

    List<DBObject> myList = new ArrayList<DBObject>();
    if (!userName.equals("all"))
        myList.add(match);
    myList.add(project);
    myList.add(group);
    myList.add(sort);
    myList.add(limit);

    List<DBObject> pipeline = myList;
    AggregationOutput output = coll.aggregate(pipeline);

Json 结果：

{ "_id" : { "projects.component" : [ "Vacation/Paid Time Off"] , "projects.client" : [ "Internal"] , "projects.name" : [ "Vacation"]} , "count" : 7}
    { "_id" : { "projects.component" : [ "Vacation/Paid Time Off"] , "projects.client" : [ "Internal"] , "projects.name" : [ "Vacation"]} , "count" : 50}
    { "_id" : { "projects.component" : [ "Daily Meeting" , "Jira-Ticket"] , "projects.client" : [ "Internal" , "Visualize"] , "projects.name" : [ "Meetings" , "Nectar Sprint 1"]} , "count" : 6}
    { "_id" : { "projects.component" : [ "Generic"] , "projects.client" : [ "BB"] , "projects.name" : [ "BBTO"]} , "count" : 45}
    { "_id" : { "projects.component" : [ ] , "projects.client" : [ "BB"] , "projects.name" : [ "BBMC"]} , "count" : 43}
    { "_id" : { "projects.component" : [ "Technology Status"] , "projects.client" : [ "BB"] , "projects.name" : [ "Meetings"]} , "count" : 42}
    { "_id" : { "projects.component" : [ "Daily Meeting" , "Nectar 1"] , "projects.client" : [ "Internal" , "Visualize"] , "projects.name" : [ "Meetings" , "Visualize"]} , "count" : 28}
    { "_id" : { "projects.component" : [ "Daily Meeting"] , "projects.client" : [ "Internal"] , "projects.name" : [ "Meetings"]} , "count" : 25}

【讨论】：