#rotation #bloom-filter #maintain #probabilistic-data-structure #item #retention #os-rng #older #twice

rotating-bloom-filter

A probabilistic data structure that rotates out old items to maintain recent membership

1 unstable release

Uses new Rust 2024

0.1.0 May 29, 2025

#1999 in Data structures

EUPL-1.2

8KB

Rotating Bloom Filter

A probabilistic data structure that maintains recent items using a dual-buffer rotation mechanism.

RotatingBloomFilter provides membership testing for recently inserted items, with older items automatically rotating out. Items are guaranteed to be retained for at least a minimum retention period, typically remaining for up to twice that duration.

Features

  • Time-limited tracking: Only maintains membership information for recently inserted items
  • Type flexible: Works with any hashable type

How it works

The filter maintains two internal Bloom filters:

  1. Current buffer: Holds items from previous rotations plus ongoing insertions
  2. Next buffer: Accumulates new insertions until rotation

When the minimum retention threshold is reached, the next buffer becomes the new current buffer, and a new next buffer is created.

Example

use rotating_bloom_filter::RotatingBloomFilter;
use rand_core::OsRng;

let mut filter = RotatingBloomFilter::new(0.01, 1000, &mut OsRng);

filter.insert("hello");
assert!(filter.contains("hello"));

// "hello" will be retained for at least 1000 insertions, but sticks around to 2000 insertions if
// the filter continues to be used.

Dependencies

~1.5MB
~33K SLoC