Time-interval series database handling in node-js and Mongodb

Question

Actually i'm creating a iot-temperature sensor database with has a event series database contain start_timestamp and end_timestamp, for every temperature change event i inserting a new document in database, My Schema is

var tempSchema = mongoose.Schema({
    sensor_name:{
        type:String
    },
    temp_value:{
        type:Number
    },
    start_timestamp:{
        type:Date,
    },
    end_timestamp:{
        type:Date,
        default:null
    }
});

for every new document insert i have to first update end_timestamp of last event and then insert new document with same timestamp which i updated last document end_timestamp.

Problems i'm facing

1) i have 30 sensors and every 2 sec i'm getting data from sensor and first i'm comparing current value with last value and if this value change i'm firing a event and update last value which has end_timestamp null,because of event loop is very busy and async nature of node sometime 2 entry of same event inserting on database the only difference start_timestamp is different (like 2020-02-22T15:52:21.639+00:00 and 2020-02-22T15:52:21.710+00:00) but ideally should be only one end_timestamp null in database

2) 2 operations is happening on same time, right now i'm using 2 function for separate opration like update_entry for end_timestamp update and post_entry for new document,is there is any method in mongodb i used to perform this operation together.

Give me your kind help Thanks in advance

Answer 1

My approach would be to create two collections (or one collection and an aggregation).

The first collection contains only the raw sensor data, I assume they have just the timestamp value:

db.collection.insertMany([
   { sensor_name: "sensor 1", temp_value: 30.2, timestamp: new Date("2020-02-22T15:52:21.639Z") },
   { sensor_name: "sensor 1", temp_value: 31.2, timestamp: new Date("2020-02-22T15:52:23.639Z") },
   { sensor_name: "sensor 1", temp_value: 32.2, timestamp: new Date("2020-02-22T15:52:25.639Z") },
   { sensor_name: "sensor 1", temp_value: 32.2, timestamp: new Date("2020-02-22T15:52:27.639Z") },
   { sensor_name: "sensor 1", temp_value: 32.2, timestamp: new Date("2020-02-22T15:52:29.639Z") },
   { sensor_name: "sensor 3", temp_value: 12.2, timestamp: new Date("2020-02-22T15:52:00.639Z") }
])

Then you can aggregate the values:

db.collection.aggregate([
   { $sort: { timestamp: 1 } },
   {
      $group: {
         _id: "$sensor_name",
         temp_value: { $last: "$temp_value" },
         start_timestamp: { $first: "$timestamp" },
         count: { $sum: 1 },
         end_timestamp: { $last: "$timestamp" },
         min_temp: { $min: "$temp_value" },
         max_temp: { $max: "$temp_value" },
      }
   }
])

If you need to store the result in a collection, append { $merge: { into: "temperatures" } } to the aggregation pipeline.