An efficient storage method of time series data

Abstract

An efficient storage method of time series data includes storing time series data collected by each data collection point separately from static attribute data, wherein each time series has a unique ID, and the corresponding time series data is stored based on the ID; storing the static attribute data separately in another database, in which each data collection point has a corresponding record inthe database. The method can save storage space, greatly improve the processing efficiency and improve the processing speed, and query processing is flexible.

Description

technical field [0001] The invention relates to the field of data processing, in particular to an efficient storage method for time series data. Background technique [0002] Time series data refers to time series data, which is a data column recorded in chronological order. A time series of data is usually generated by a data collection point (sensor or device), and this collection point often has various attributes, such as device serial number, manufacturer, model, geographical location, etc., which are static information . [0003] In actual scenarios, there are often multiple devices of the same type, and the time series data generated by them need to be aggregated and calculated based on various attributes. For example, to calculate the average value of PM2.5 in Beijing, you need to add up the values ​​of each PM2.5 monitoring point in Beijing to find the average, and to calculate the average value of PM2.5 in Chaoyang District, you only need to add up the values ​​o...

AI Technical Summary

Problems solved by technology

But this design brings two problems: data redundancy is serious, because each record has a label; and it is very difficult to modify the label information of historical data, resulting in insufficient query flexibility

Images