v.21.12New Features

Introduced More Effective Exponential Time Decayed Window Functions

Official CH changelog text
Introduced window functions exponentialTimeDecayedSum, exponentialTimeDecayedMax, exponentialTimeDecayedCount and exponentialTimeDecayedAvg which are more effective than exponentialMovingAverage for bigger windows. Also more use-cases were covered. #29799 (Vladimir Chebotarev).
Introduced new window functions exponentialTimeDecayedSum, exponentialTimeDecayedMax, exponentialTimeDecayedCount, and exponentialTimeDecayedAvg which provide more efficient and accurate analysis over larger windows compared to exponentialMovingAverage.

Why it matters

These new window functions address the limitations of exponentialMovingAverage when applied to bigger windows by offering better performance and wider analytical use-cases, enabling users to perform decay-based aggregations more effectively on time-series data.

How to use it

Use the newly introduced window functions within your SELECT queries over window frames. Replace or supplement exponentialMovingAverage with exponentialTimeDecayedSum, exponentialTimeDecayedMax, exponentialTimeDecayedCount, or exponentialTimeDecayedAvg for improved performance and analysis over larger data sets.

Related resources

PreviousNext