On Wed, 21 Dec 2005, Jim Gray wrote:
> There is also the cross match that Tanu & Maria implemented in SQL.
> That may be the easiest way to do the implementation if the nodes all
> have a SQL backend.
> When things are "in" the database it runs fast ~100k object/sec for
> small radius.
> But, getting objects into a node dominates all other costs (moving stuff
> thru xml is expensive).
> We recently found some logic errors in our "zones" cross match (near the
> meridian and poles) but seem to have a fix for them now.
> Maria can tell you all about this (it is the subject of her PhD).
Jim (and lurkers),
certainly the STIL/STILTS match is not going to compete with an SQL one which makes use of precomputed indices; it's designed for the case where you've got data with no advance preparation (typically just an unsorted list of (RA,Dec,other-data) rows, for instance a FITS table or VOTable streamed or on disk). Since Pedro says he's interested in client-side matching, I presume that the data's not going to be in a database and in general no RDBMS will even be available for use. However, it depends a bit on what is meant here by "client".
Mark
-- Mark Taylor Astronomical Programmer Physics, Bristol University, UK m.b.taylor@bris.ac.uk +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/Received on 2005-12-21Z15:42:54