Comments on: Bloom Filters Quickie

By: Felix Contrera

Felix Contrera — Mon, 03 Oct 2011 06:43:04 +0000

It’s perfect time to make some plans for the future and it is time to be happy. I’ve read this post and if I could I wish to suggest you few interesting things or tips. Perhaps you could write next articles referring to this article. I want to read even more things about it!

By: Ti4iT

Ti4iT — Tue, 14 Apr 2009 09:53:34 +0000

thank you for this awsome work

By: Andrei Zmievski’s Blog: Bloom Filters Quickie : Dragonfly Networks

Andrei Zmievski’s Blog: Bloom Filters Quickie : Dragonfly Networks — Wed, 08 Apr 2009 04:51:24 +0000

[...] Zmievski has written a new post about a new extension he’s worked up (out of curiosity for the technology) – the pecl/bloomy [...]

By: Andrei Zmievski’s Blog: Bloom Filters Quickie : WebNetiques, LLC : Website Developers in Minneapolis, MN

Wed, 08 Apr 2009 04:48:51 +0000

[...] Zmievski has written a new post about a new extension he’s worked up (out of curiosity for the technology) – the pecl/bloomy [...]

By: 網站製作學習誌 » [Web] 連結分享

網站製作學習誌 » [Web] 連結分享 — Wed, 08 Apr 2009 02:53:34 +0000

[...] Bloom Filters Quickie [...]

By: Wolfram Hüsken » Bloomfilter mit PHP

Wolfram Hüsken » Bloomfilter mit PHP — Mon, 06 Apr 2009 11:56:20 +0000

[...] Nachdem Andrei Zmievski am WE eine kleine Bloomfilter Erweiterung in PECL online gestellt hat, habe ich mir ein mögliches Einsatzgebiet dafür gesucht: [...]

By: Andrei

Andrei — Sat, 04 Apr 2009 08:21:46 +0000

@Ryan, sorry, you’re correct. Thanks for your note.

By: Ryan

Ryan — Sat, 04 Apr 2009 08:05:53 +0000

One of your statements is incorrect. You state:

“The $answer2 will always be “no” because “zoo” was never added to the filter.”

This is incorrect. It is possible that “zoo” could return a ‘yes’ even if it hadn’t been added yet. This is because its possible it hashes to the same values as a previously added string. In your example, if ‘zoo’ hashed to the same values as ‘foo’ or ‘bar’, then this would happen.

However, A Bloom Filter will never return a ‘no’ when the element IS in the filter. A BF can definitely return a yes both when an element is in there, or not in there. This has everything to do with the false positive probability.

By: Andrei

Andrei — Fri, 03 Apr 2009 22:33:24 +0000

@Matt: right now it supports only strings, so essentially any data type that can be represented in the string form. Is that too much of a limitation?

@Pablo: BloomFilter class can serialize itself to a string, so you can simply do serialize($filter) and store it in APC or memcache or elsewhere.

By: thejo

thejo — Fri, 03 Apr 2009 22:30:00 +0000

This is excellent, but like Pablo mentioned, persistence of the data structure is important in many situations. For example, did I already receive a request with this combination of parameters?