Fundamentals of Apache IoTDB UDF and its practice in the field of industrial data quality

苏宇荣,贺文迪

Chinese Session 2022-07-30 18:10 GMT+8  #iot

Apache IoTDB UDF framework provides a set of easy to use, semantically rich Java API for users. Users can quickly build powerful timeseries processing logic based on this framework: The original data can be consumed according to: Point by Point, Sliding Time Window, Sliding Count Window and Session Window, so as to realize the aggregation, deformation or amplification of Timeseries. At present, the community has implemented more than 100 functions based on this framework. With the nested expression feature of Apache IoTDB, the service burden in timeseries analysis scenarios can be greatly reduced.

This presentation will focus on the Apache IoTDB UDF framework and consists of two main parts:

  1. Basic knowledge of Apache IoTDB UDF: including the introduction of UDF interface semantics, programming examples, usage examples and future prospects, etc.

  2. Practice of Apache IoTDB UDF in the field of industrial data quality: including the introduction of data portrait, data repair, data matching, anomaly detection, sequence discovery and other functions.

Speakers:


Yurong Su: School of Software, Tsinghua University, Master student, master candidate at the School of Software, Tsinghua University. His mainly interests include distributed systems, timeseries databases, and streaming computing engines. Open source enthusiast, Apache IoTDB Committer. He focused on improving the data processing capabilities of IoTDB and promoted the implementation of IoTDB functions (UDF) framework, expression framework, triggers, continuous queries and other features in the community.
Wendi He: School of Software, Tsinghua University, Master student, School of Software, Tsinghua University, Apache IoTDB Contributor. Core developer of the Apache IoTDB-Quality project, focusing on the IoTDB-Quality user custom functions part of the work.