W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2004

Re: N-Triples Parser for Python

From: Phil Dawes <pdawes@users.sf.net>
Date: Wed, 20 Oct 2004 23:37:25 +0000
Message-ID: <16758.63157.374419.448322@gargle.gargle.HOWL>
To: Chris Purcell <cjp39@cam.ac.uk>
Cc: www-rdf-interest@w3.org

Hi Chris,

Chris Purcell writes:
 > Just out of curiosity, what are you defining as "fast"? On my server, I 
 > found MySQL peaks at around 2K (simple) transactions a second -- 
 > sufficient for some purposes, laughably poor for others. Is there 
 > anything smart you've done to get around this?
 > 

I'm afraid not - I'm defining fast as 'fast enough for my needs' ;-) 
I haven't done any volume tps studies, and my loads are really low.

Most triple-pattern-match queries (even reasonably complicated ones)
come in at a few 10s to a few 100s of milliseconds. 
Label search queries are slow because they rely on a regex table scan
- I'd be interested in strategies to speed this up (maybe a python
lucene equivalent[1]).  I use veudas for interactive knowledge
management servicing a handful of people, so anything under half a
second is fine for me.

Sorry about that - I feel a bit silly saying 'fast' now. I'd be keen
to benchmark it if anybody has some example data + queries to try.

Cheers,

Phil

[1] http://www.divmod.org/Home/Projects/Lupy/
Received on Wednesday, 20 October 2004 22:35:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:10 GMT