Discussion on: How does database indexing work?

louy2 profile image
Yufan Lou

To my knowledge mainstream relational databases (MySQL, PostgreSQL, SQLite) all use a form of B-tree, because hashtable is not good for comparison. Abstractly you can view B-tree as a sorted set while a hashtable an unordered set. For string operations I suspect they are just implemented as regular expressions, which if you limit them to no look back can be implemented efficiently as finite state machines.