mongodb聚合函数操作

今天碰到需要统计热门标签，也就是标签名访问次数，用sql很简单

select lable_id,count(*) as count from table where state>-1 group label

但是mongodb显然没那么轻松，故在网络上收集资料，本地测试，最终解决了问题

代码如下：

<span style="font-size:14px;"><span style="white-space:pre">	</span>   DBCollection myColl = db.getCollection("toskebre");

//        group 字段
            DBObject groupFields = new BasicDBObject( "_id", "$label_id");
            groupFields.put("count", new BasicDBObject( "$sum", 1));

            DBObject group = new BasicDBObject("$group", groupFields );

//            倒序排列
            DBObject sortFields = new BasicDBObject("count", -1);
            DBObject sort2 = new BasicDBObject("$sort", sortFields ); 
            //去前10条记录
            DBObject limitFields = new BasicDBObject("$limit", 10);
            //state>0
            DBObject filterFields = new BasicDBObject("$gt", -1);
            DBObject filter2 = new BasicDBObject("state", filterFields);
            DBObject filter = new BasicDBObject("$match", filter2);

//           获取数据
            AggregationOutput output =  myColl .aggregate(filter,group, sort2,limitFields);
            Iterable<DBObject> outputResults = output.results();</span>

DBObject的value参数必须是DBObject

函数DBCollection.aggregate(DBObject, DBObject...)，参数可以是List<DBObject>也可是多个 DBObject参数，

结果：

[{ "_id" : { "$oid" : "5668e49da8e24d0170be62f5"} , "count" : 10}, { "_id" : { "$oid" : "5668e410a8e24d0170be62f3"} , "count" : 2}]

希望对大家有点帮助，错误之处还请指出

主要参考：http://blog.csdn.net/miyatang/article/details/20997313

这个地址也有其他详细的sql聚合对应操作和例子