Section: New Software and Platforms

BBhash

Basic binary representation of successive hash

Keywords: C++ - Indexation - Data structures

Functional Description

BBHash is a simple library for building minimal perfect hash function. Given a set of N input keys, it will compute a bijective function that will associate to each key an integer between 1 and N. This then allows to create an indexed array that will hold some data for each key. It is designed to handle large scale datasets (hundred billion and more elements). The function itself is just a little bit larger than other state-of-the-art libraries, it takes approximately 3 bits / elements (compared to 2.62 bits/elem for the emphf lib), but construction is faster and does not require additional memory.

• Participants: Guillaume Rizk, Pierre Peterlongo, Rayan Chikhi and Antoine Limasset

• Contact: Guillaume Rizk

• URL: https://github.com/rizkg/BBHash