[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: A Data Reduction Proposal




Chris,

For matching approximately 200,000 stars, using PostgreSQL 7.2b2 (and
vacuuming every 10,000 stars processed, maybe a little excessive) on an AMD
K6-2 500MHZ, 640Mb RAM, ATA-IDE disks (slow machine by today's standards),
it took me ~4 hours.  I can get exact timings and such when I return home
tonight.  I believe inserting the records (not copying) took about 1 hour.

A little slow, but doable for a single site.  I figure a new(er), server
style machine might be able to keep up, especially if only 'interesting', or
a subset of data is sent after initial crunching on the local machine.

I'd be happy to dump the dB and send out a CD.  Or, I can send out (e-mail)
the schema, scripts, trigger and star lists (you'll have to supply the raw
Tycho2 catalog, which a script will populate the dB with).  Note that the dB
is in rough form as I wait till I have something working before I go back
and clean it up.  Notably, I use a trigger to update the 'master' tables,
but I have not added a pre-delete rule so the tables cascade correctly.

Later,
Rob

> -----Original Message-----
> From: Chris Albertson [mailto:chrisalbertson90278@yahoo.com]
> Sent: Monday, February 04, 2002 9:59 AM
> To: Creager, Robert S; tass@listserv.wwa.com
> Subject: RE: A Data Reduction Proposal
> 
> 
> I tried that a while back and got very poor performance and
> gave up.  That was usiing Postgresql 6.x  How fast are you
> matches going?  What size data sets and could you describe the
> computer you use?
> 
> This is really the best way to go if it works for TASS Mk IV
> sized data but it was taking _days_ with Mk III data so I did
> it in pure "C".
> 
> 
> 
> --- "Creager, Robert S" <CreagRS@LOUISVILLE.STORTEK.COM> wrote:
> > 
> > If anyone is interested, I have some Perl/SQL (PostgreSQL) 
> which does
> > #5 in
> > Tom's list below for a single radius (.004 degrees is what I
> > currently use
> > along with at least 10 measurements).  Basically, all the radius
> > matching is
> > accomplished using PostgreSQL's circle_contain_pt and using a
> > bounding box
> > around the circle to gain index usage to minimize the stars to
> > search.
> 
> 
> =====
> Chris Albertson 
>   Home:   310-376-1029  chrisalbertson90278@yahoo.com
>   Cell:   310-990-7550
>   Office: 310-336-5189  Christopher.J.Albertson@aero.org
> 
> __________________________________________________
> Do You Yahoo!?
> Great stuff seeking new owners in Yahoo! Auctions! 
> http://auctions.yahoo.com
>