semi-perfect hashing

carr@e.gandalf.ca (Dave Carr)
Wed, 20 Jun 90 04:05:58 GMT

          From comp.compilers

Related articles
semi-perfect hashing carr@e.gandalf.ca (1990-06-20)
| List of all articles for this month |

Newsgroups: comp.arch,comp.compilers,comp.databases
From: carr@e.gandalf.ca (Dave Carr)
Date: Wed, 20 Jun 90 04:05:58 GMT
Organization: Compilers Central
Keywords: hashing perfect

I am looking for information on semi-PERFECT HASH routines.
Specifically, I require:


(1) A maximum collision chain length of 4 (minimal hash not required);
(2) A spread factor of 32:1. Perfection not required.
        The loading of the hash table will be only 3 percent.


The hash must operate on fixed length character strings of 8 characters.


An algorithm on order N or NlogN complexity is required.


I have already examined the GNU GPERF Utility. It seems a bit complex
for what I require. Also, since it does not iterate, it does not
garauntee a solution.


The code generator is not required, only the algorithm.


Any help will be greatly appreciated. Please respond by E-mail.


Dave Carr | carr@e.gandalf.ca
Gandalf Data Limited | TEL (613) 723-6500
Nepean, Ontario, Canada | FAX (613) 226-1717
[Every perfect hash generator that I've seen is very slow. -John]
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.