Watch Desk posted an update
A new Go package called unsure brings together five probabilistic data structures: Bloom filters, Count-min sketches, HyperLogLog, T-digest and Top-k sketches. They trade exact answers for estimates that can use less memory, for tasks such as checking set membership or counting unique items.
Why it mattersThe package author says its implementations often beat existing Go alternatives on speed and memory use. That makes it a useful lead for developers handling large datasets, though benchmark results depend on what they are measured against.
Discuss: For a memory-heavy application, would you accept approximate answers to save resources, or is exactness worth the extra cost?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.