Implementing Page UV Counting with Redis Set, Hash, Bitmap, and HyperLogLog Using Redisson

In mobile internet scenarios with massive data, we often need to associate a key with a collection and perform cardinality statistics such as daily active users (DAU), unique visitors (UV), search term counts, and registration IP counts.

The article demonstrates four Redis data structures—Set, Hash, Bitmap, and HyperLogLog—to implement UV counting for a page, gradually introducing each method and its trade‑offs.

Using Set

Because a user visiting multiple times in a day should be counted once, a Redis Set can store unique user IDs. Example commands:

SADD Redis为什么这么快:uv 肖菜鸡 谢霸哥 肖菜鸡
SCARD Redis为什么这么快:uv

SCARD returns the number of unique IDs.

Using Hash

Store each user ID as a hash field with value 1; HLEN gives the UV count.

HSET Redis为什么这么快 肖菜鸡 1
HLEN Redis为什么这么快

Using Bitmap

Bitmap uses a bit array where each bit represents a user. SETBIT writes a bit, BITCOUNT counts bits set to 1. User IDs must be converted to numeric offsets.

SETBIT 巧用Redis数据类型实现亿级数据统计 6 1
BITCOUNT 巧用Redis数据类型实现亿级数据统计

HyperLogLog

HyperLogLog provides probabilistic cardinality estimation with fixed 12 KB memory regardless of set size, suitable for millions‑to‑billions of elements. Commands PFADD, PFCOUNT, and PFMERGE are used.

PFADD Redis主从同步原理:uv userID1 userID2 userID3
PFCOUNT Redis主从同步原理:uv
PFMERGE database Redis数据 MySQL数据
PFCOUNT database

Redisson Integration

Java code shows how to add elements, addAll, merge, and count using Redisson's RHyperLogLog API.

public <T> void add(String logName, T item) {
    RHyperLogLog<T> hyperLogLog = redissonClient.getHyperLogLog(logName);
    hyperLogLog.add(item);
}

public <T> void addAll(String logName, List<T> items) {
    RHyperLogLog<T> hyperLogLog = redissonClient.getHyperLogLog(logName);
    hyperLogLog.addAll(items);
}

public void merge(String logName, String... otherLogNames) {
    RHyperLogLog<T> hyperLogLog = redissonClient.getHyperLogLog(logName);
    hyperLogLog.mergeWith(otherLogNames);
}

public long count(String logName) {
    RHyperLogLog<T> hyperLogLog = redissonClient.getHyperLogLog(logName);
    return hyperLogLog.count();
}

Summary

Set offers simple deduplication but high memory for large cardinalities; Hash gives high precision for moderate data; Bitmap is efficient for binary‑state statistics; HyperLogLog achieves near‑constant memory with acceptable error (~0.81 %).