Table design for multiple time series data

问题

I have more than 200 separate time series data(each represent one variable) that I gather from different sources/REST API calls.

The frequency of each variable is different. Example temperature data is coming at very high frequency, but status data is very less frequent.

I am looking for suggestions for scalable table design to store these data. If I store all the data in one table with timestamp being the key, I think the table will have so much nulls.

回答1:

Based on your description, my first thought is something like this:

Create Table Data_Type
(
    ID Int Identity
    , Data_Type_Description VarChar(100)
)

Create Table Data_Values
(
    ID Int Identity
    , Data_Value_Time_Stamp TimeStamp
    , Data_Type_ID Int       -- foreign key to Data_Type
    , Value Numeric(17, 4)   -- I'm guessing here
)

Does that make sense?

来源：https://stackoverflow.com/questions/54587600/table-design-for-multiple-time-series-data