Timestamp as Key in mongodb as timeseries database

Question

I want to use mongodb as timeseries database and query via timestamp + id.

Mongodb as shown a way to store data here.

{
  timestamp_hour: ISODate("2013-10-10T23:00:00.000Z"),
  ID: “System1”,
  values: {
    0: { 0: 999999, 1: 999999, …, 59: 1000000 },
    1: { 0: 2000000, 1: 2000000, …, 59: 1000000 },
    …,
    58: { 0: 1600000, 1: 1200000, …, 59: 1100000 },
    59: { 0: 1300000, 1: 1400000, …, 59: 1500000 }
  }
}

but i have multiple values and for every value a timestamp, its not periodic. Data has some time a delay or is not coming for days. So i dont want to use 0-24 for my hours and 0-59 for my minutes. Can i use instead my measured timestamp? every value in my document has the same timestamp, so if value1 has 50 entries , value2 has also 50 entries and equal timestamp.

     {
          timestamp_hour: ISODate("2013-10-10T23:00:00.000Z"),
          ID: “System1”,
          values1: {
            "2013-10-10T22:00: {
         "2013-10-10T22:01:00.000Z": 999999, 
         "2013-10-10T22:02:00.000Z": 999999,
         "2013-10-10T22:03:00.000Z": 1000000 
                                },
            "2013-10-10T23:00:": { 
         "2013-10-10T23:01:00.000Z": 2000000,
         "2013-10-10T23:02:00.000Z": 2000000, 
                                 },
          }
         values2: {
            "2013-10-10T22:00: {
         "2013-10-10T22:01:00.000Z": 999999, 
         "2013-10-10T22:02:00.000Z": 999999,
         "2013-10-10T22:03:00.000Z": 1000000 
                                },
            "2013-10-10T23:00:": { 
         "2013-10-10T23:01:00.000Z": 2000000,
         "2013-10-10T23:02:00.000Z": 2000000, 
                                 },
         }
    }

score 2 · Accepted Answer · edited May 23 '17 at 11:46

You could store your values with the incoming timestamp, and use a separate mongodb map-reduce process to transform your values and store them in different collections with the granularity you need (ex. hourly, daily, monthly). See also incremental map-reduce.

You could also find these posts inspiring:

score 0 · Answer 2 · answered Mar 02 '22 at 07:10

It's a pretty poor choice. Of course you can store data like this. But when the data volume is huge, like huge number of data collection points with each have huge number of data (for example, the interval is 10 seconds), MongoDB will become a nightmare.

Try to use real timeseries DB for timeseries data.

Timestamp as Key in mongodb as timeseries database

2 Answers2