mongodb储存即时聊天的时候如何取出用户聊天列表

Question

{
        "_id" : ObjectId("576cfd363325ffaa1dbdde15"),
        "current_uid" : "5",
        "to_uid" : "3",
        "content" : "你大爷"
    },
    {
        "_id" : ObjectId("576cfd6e3325ff501e07f4ae"),
        "current_uid" : "5",
        "to_uid" : "3",
        "content" : "郭德纲"
    },
    {
        "_id" : ObjectId("576cfe753325ff501ea76603"),
        "current_uid" : "5",
        "to_uid" : "4",
        "content" : "325235"
    }

存的时候如上所示，想要获取用户ID为5的这个跟哪些人聊过天，也就是3和4这2个用户，现在想要获取包括content的信息，也就是跟该用户最新的一条记录，比如3这个用户，我要获取的就是郭德纲这个content而不是上面那条，这个应该怎么查询，初学mongodb，球帮助

Answer 1

This is a very typical groupingTOP Nproblem. There are also corresponding expressions in SQL databases, such as SQLServer

ROW_NUMBER() OVER(PARTITION BY COLUMN ORDER BY COLUMN)

Using aggregation就可以解决了。aggregation in MongoDB is a big topic. It is impossible to explain all the content here. You can only check the documentation to understand the various operators used below.

db.chat.aggregate([
    {$match: {current_uid: "5"}},
    {$sort: {_id: -1}},
    {$group: {_id: {current_uid: "$current_uid", to_uid: "$to_uid"}, content: {$first: "$content"}}}
])

A few explanations:

1. Use {$sort: {_id: -1}}其实就是时间倒序。_id which contains time, and sorting it can most of the time be considered as sorting time;

2. $first取到了第一个元素。如果想取头n个元素怎么办？$push+$slice That’s it;

3. How can the above query be made faster? db.chat.createIndex({current_uid: 1, _id: 1})