How to Track Active Users with Redis Bitmaps: A Step‑by‑Step Guide

Scenario

When a user logs in, the login must be recorded for later statistical analysis. The main requirements are:

Number of users logged in today.

Users who have logged in within the last three days.

Users who have logged in within the last seven days.

Conventional Solution

Without Redis, the typical approach is to write a log entry at each login or insert a record into a database table, then run scheduled jobs to aggregate the data according to the requirements.

Redis Solution

Redis bit operations are ideal for this scenario because each bit can represent a binary state (0 or 1) indicating whether a specific user logged in on a given day. By storing a day's login information in a single key, each offset corresponds to a user ID.

Basic Bit Operations Example

(1) Set the value of a specific offset in a key to 0 or 1: setbit key 100 1 (2) Bitwise operations:

For example, given:

key1 -> 0101
key2 -> 0011

AND operation: bitop and ret key1 key2 Result stored in ret is 0001.

OR operation: bitop or ret key1 key2 Result stored in ret is 0111.

(3) Count bits set to 1:

bitcount key

Applying the Solution

Assume today is 2016‑01‑18 and the key is named userlogin:20160118.

(1) User login: setbit userlogin:20160118 100 1 (2) Count users logged in today: bitcount userlogin:20160118 (3) Count users who logged in all of the last three days (bit value 1 in every day):

bitop and ret userlogin:20160116 userlogin:20160117 userlogin:20160118

(4) Count users who logged in at least once in the last seven days (any bit set to 1):

bitop or ret userlogin:20160112 userlogin:20160113 userlogin:20160114 userlogin:20160115 userlogin:20160116 userlogin:20160117 userlogin:20160118