Time Series Data Management via Location Index Table
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for querying time series data are inefficient due to the need to check large data tables row by row, leading to low query efficiency and increased system load.
Innovation Solution
A time series data management method that uses a target data table and a location index table to store and query time series data, where only the latest data is stored in the target table and previous data is stored in separate blocks, allowing direct retrieval from the index table without row-by-row checking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If time series data is stored in a data table with each row representing a time period, then all time series data can be stored, but query efficiency deteriorates due to row-by-row checking
Solution Approach 1:
The patent segments time series data storage by creating separate storage locations for different time periods. Instead of storing all data in a single sequential table, the system divides data into multiple segments (first time period data, second time period data, etc.) stored in different locations, with an index table tracking where each segment is stored. This allows queries to jump directly to relevant segments rather than checking every row sequentially.
Solution Approach 2:
The patent introduces an index table as an intermediary between the query request and the actual time series data storage. The index table stores mapping relationships between time periods and their corresponding storage locations, acting as a mediator that translates query requirements into direct access paths. This intermediary eliminates the need for row-by-row scanning by providing direct pointers to data locations.
2Reliability
If the data table is checked from the first row to the last row to query time series data, then all data can be accessed, but the time consumed increases significantly
Solution Approach 1:
The patent performs preliminary action by pre-organizing time series data into distinct time period segments and pre-establishing an index table that maps each time period to its storage location. This preliminary organization occurs during data ingestion, so when queries arrive, the system already has a ready-made roadmap to the data locations, eliminating the need for time-consuming sequential searches while ensuring all required data is accessible.
3Productivity
If a location index table is introduced to store storage locations of time series data, then query efficiency improves through direct access, but device complexity increases
Solution Approach 1:
The patent applies universality by designing the index table to serve multiple functions simultaneously: it stores storage location mappings, tracks time period associations, and provides query routing information. This multi-functional index structure reduces the need for separate data structures for each function, thereby limiting the increase in overall system complexity while achieving efficient direct access to time series data.
Data Source
AI summary
This application discloses a time series data management method, device, and apparatus. A device receives a data query request, wherein the data query request includes an identifier of the first object. The device obtains first time series data corresponding to the first object in a target data table according to the identifier of the first object. The device determines a second storage location that is in a location index table and that is stored in a first storage location corresponding to the first object. Further, the device obtains second time series data of the first object and a third storage location corresponding to the second time series data that are stored in the second storage location, The device also obtains other time series data corresponding to the first object according to the third storage location.