Mongo对象子集合聚合统计，和对应Java实现

统计子集合里面的字段，数据格式如下（表名是DebutResult）：

{
    "_id" : ObjectId("5fa59607b754a52813412722"),
    "taskId" : "5fa58f60b754a569b083d8ac",
    "imsi" : "460110730176756",
    "regional" : "四川省自贡市",
    "firstPlaceId" : "5c505e17498e4a19f0007cbd",
    "firstPlaceName" : "星维电影院",
    "firstCatchTime" : NumberLong(1603411840),
    "catchPointList" : [ 
        {
            "placeId" : "5c505e17498e4a19f0007cbd",
            "catchTime" : NumberLong(1603411840)
        }, 
        {
            "placeId" : "5c74bb5d498ef85d41240622",
            "catchTime" : NumberLong(1603412001)
        }, 
        {
            "placeId" : "5c505e17498e4a19f0007cb9",
            "catchTime" : NumberLong(1603411949)
        }
    ],
    "cdt" : ISODate("2020-11-07T02:29:27.929+08:00")
}

现在要统计指定taskId下面，所有的抓取场所（catchPointList.placeId）和 对应抓取的次数

最后的mongo语句如下：

db.debutResult.aggregate([

{$match: {"taskId": "5fa9f528b754a53776db15ef"}},

{$project: {"_id": 0, "catchPointList": 1}},

{$unwind: "$catchPointList"},

{$group: {_id: "$catchPointList.placeId", countResult: {$sum: 1}}},

{$project: { "_id": 0, "countResult" : 1, "placeId" : "$_id" }}

])

最终的结果是这样的：

 

 

 下面简单解释一下每条的大概意思：

db.debutResult.aggregate([

//过滤条件
{$match: {"taskId": "5fa9f528b754a53776db15ef"}},

//字段转换，保留有用到的字段
{$project: {"_id": 0, "catchPointList": 1}},

//对子集合的处理
{$unwind: "$catchPointList"},

//类似groupBy，分组统计
{$group: {_id: "$catchPointList.placeId", countResult: {$sum: 1}}},

//最后对输出字段再转换一次，方便Java代码对象接收
{$project: { "_id": 0, "countResult" : 1, "placeId" : "$_id" }}

])

其中，$unwind处理后，输出的数据是这样的：

 

 

 

下面介绍一下Java里面代码是什么样的：

　　@Override
    public List<PlaceCountBO> getCatchPointByTaskId(String taskId) {

        Criteria where = Criteria.where("taskId").is(taskId);
        MatchOperation match = Aggregation.match(where);

        //可以省略
        ProjectionOperation project = Aggregation.project("catchPointList");

        UnwindOperation unwind = Aggregation.unwind("catchPointList");

        GroupOperation group = Aggregation.group("catchPointList.placeId").count().as("countResult");

        ProjectionOperation project2 = Aggregation.project("countResult").and("_id").as("placeId");

        Aggregation agg = Aggregation.newAggregation(match, project, unwind, group, project2);

        //log.info("DebutFollowResult统计agg = {}", agg.toString());
        AggregationResults<PlaceCountBO> results = mongoTemplate.aggregate(agg, DebutResult.class, PlaceCountBO.class);
        List<PlaceCountBO> countBOList = results.getMappedResults();

        return countBOList;
    }

其中，PlaceCountBO对象里面，只要包含"countResult"和“placeId”字段就行了，可以接收聚合返回的结果

最后，Java环境是spring boot的，引入有关的包就可以了

　　　　　<dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-data-mongodb</artifactId>
        </dependency>