Redis 4.x Cookbook

Using the HyperLogLog data type

Counting distinct values is a common task in various kinds of daily data processing scenarios. In Redis, while it's fine and sometimes necessary to implement distinct counting using set, memory consumption and performance degradation should be taken into consideration when the size of the set increases to tens of millions. If you don't need to retrieve the content of the data set and just want a unique counting value, one thing you can do is to use the HyperLogLog (HLL) data type in Redis to optimize the memory and performance issues caused by the set. In this recipe, we'll cover how to use HLL in Redis.