Yes, that would be berkely style DBM databases. If you're running on Linux (and who isn't? 🙂 Then you can use the --with-gdbm switch to support this.
Basically this gives you the ability to put your data into very fast, very simple hash style lookup tables.