On 1/27/2016 9:49 AM, scott coston wrote:
yes, me neither (11pm for me). i just rechecked my spreadsheet and i typed in
128 where i should have typed 256. but it doesn't really matter as the my test
file is small compared to where the file will be going.
Mine was worse - it's downright embarrassing. :-[ When I was running
the numbers somehow I put in 6e10 as the the scale - 60 billion
instead of the 6 million you specified - and I didn't realize it until
after I had sent the reply.
Your analysis of the hash size is quite reasonable (modulo allocator
overhead). I do think, though, that you can save considerable space
with a byte array vs pointer array + bignums.
anyway, since you seem to have some ideas i have more info for you. take a look
at this:
https://www.aaai.org/Papers/AAAI/1996/AAAI96-178.pdf
I'll take a look at it.
I haven't had to figure out an external merge filter for decades. It can
be tricky when the number of files needs to be variable. And also as I
mentioned, the fact that you want it sorted on the non-filtered field is
a complication.
________________________________________
From: racket-users@googlegroups.com <racket-users@googlegroups.com> on behalf of
George Neuner <gneun...@comcast.net>
Sent: Wednesday, January 27, 2016 4:28 AM
To: racket-users@googlegroups.com
Subject: Re: [racket-users] Re: appending files
Sorry. I shouldn't do math at 4am. Ignore the numbers. However, it
is still correct that the byte array will use less space than an array
of bignums.
George
--
You received this message because you are subscribed to the Google Groups "Racket
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.