Send Beginners mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
http://www.haskell.org/mailman/listinfo/beginners
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Beginners digest..."
Today's Topics:
1. Re: attoparsec and EOF (Arthur Clune)
2. Re: attoparsec and EOF (Arthur Clune)
3. Re: histogram over large data (Rados?aw Szymczyszyn)
4. Re: histogram over large data (Stephen Tetley)
5. Re: histogram over large data (Ian Knopke)
6. graphics.gloss errors (Gregory Guthrie)
7. wxHaskell install errors (Gregory Guthrie)
8. Re: graphics.gloss errors (Brent Yorgey)
9. Re: graphics.gloss errors (Gregory Guthrie)
10. Re: histogram over large data (Stephen Tetley)
----------------------------------------------------------------------
Message: 1
Date: Tue, 5 Jun 2012 15:29:03 +0100
From: Arthur Clune <[email protected]>
Subject: Re: [Haskell-beginners] attoparsec and EOF
To: [email protected]
Message-ID:
<CAAa4kjx1T=oMqhFL+a4vzxR61nUea0ETWP+iyX=ozkqnj8o...@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
And to answer my own question.....I should use parseOnly.
--
Arthur Clune [email protected]
On Tue, Jun 5, 2012 at 10:56 AM, Arthur Clune <[email protected]> wrote:
> I'm writing a simple attoparsec parser, which works fine except for
> handling end of file:
------------------------------
Message: 2
Date: Tue, 5 Jun 2012 15:42:55 +0100
From: Arthur Clune <[email protected]>
Subject: Re: [Haskell-beginners] attoparsec and EOF
To: [email protected]
Message-ID:
<CAAa4kjw0UpB9SFDfGWThaQRPbVW-w7obELuY=qe=dkiubfr...@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
And to answer my own question.....I should use parseOnly.
--
Arthur Clune [email protected]
On Tue, Jun 5, 2012 at 10:56 AM, Arthur Clune <[email protected]> wrote:
> I'm writing a simple attoparsec parser, which works fine except for
> handling end of file:
------------------------------
Message: 3
Date: Tue, 5 Jun 2012 20:41:42 +0200
From: Rados?aw Szymczyszyn <[email protected]>
Subject: Re: [Haskell-beginners] histogram over large data
To: [email protected]
Message-ID:
<CAG=dco3qq++adaxfrytifr0r5qfesc3gsjhghiefzabsheu...@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Hi Ian,
In case you were looking for an example to get your teeth into you
might be interested in these: https://gist.github.com/2876666
These two scripts both serve the same purpose of building a map of
word counts from a text file. They both use Data.Text for Unicode IO,
but each tests a different structure. Though unordered-containers
package with its Data.HashMap is often suggested as an efficient
mapping structure, in my case (of these two scripts) the
Data.HashTable from standard library wins taking circa half the time
to run on the same dataset (though it's not purely functional as its
actions operate in the IO monad).
Finally, what puzzles me the most, is that a roughly equivalent script
in Python which just reads the same datafile into a standard dict
performs in about 1/3 of the time of the faster one of the above two
and Python's hardly a fast language... Bewildering, indeed.
Hope I didn't put you off :)
------------------------------
Message: 4
Date: Tue, 5 Jun 2012 21:06:00 +0100
From: Stephen Tetley <[email protected]>
Subject: Re: [Haskell-beginners] histogram over large data
To: Rados?aw Szymczyszyn <[email protected]>
Cc: [email protected]
Message-ID:
<CAB2TPRBR9c5b3X5K7aaaHp5sOgJ8TZUeEnkmqSjU83zmJ2f=l...@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
On 5 June 2012 19:41, Rados?aw Szymczyszyn <[email protected]> wrote:
>
> Finally, what puzzles me the most, is that a roughly equivalent script
> in Python which just reads the same datafile into a standard dict
> performs in about 1/3 of the time of the faster one of the above two
> and Python's hardly a fast language... Bewildering, indeed.
Dicts are efficient in Python though (as they are efficiently
implemented in C). Python often seems to beat Haskell in
micro-benchmarks that just fill a dictionary then run a simple query /
summation on it.
------------------------------
Message: 5
Date: Tue, 5 Jun 2012 21:09:49 +0100
From: Ian Knopke <[email protected]>
Subject: Re: [Haskell-beginners] histogram over large data
To: Stephen Tetley <[email protected]>
Cc: [email protected]
Message-ID:
<cac+f4wmk3kovxddjqth79w8diftaqufm3azovlnxmk25pme...@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Hi Stephen,
I see now I didn't explain myself especially well. One way to build a
frequency count in haskell comes from LYAHFGG, chapter 7:
ghci> map (\l@(x:xs) -> (x,length l)) . group . sort $
[1,1,1,1,2,2,2,2,3,3,2,2,2,5,6,7]
[(1,4),(2,7),(3,2),(5,1),(6,1),(7,1)]
my guess would be that the problem with this, unless haskell is a lot
more magical than I thought, is that it requires all the entries to be
held in memory to do the sort (and then the group I suppose). The
database has hundreds of millions of entries.
My previous example in perl builds results cumulatively as it
retrieves each entry from the db - get_next_from_db() is basically an
iterator that returns undef once it runs out of entries (using DBI,
which is an SQL wrapper very similar to HDBC. I only wrote it that way
to focus on the real problem.) I could also do this in haskell with
something like accumArray but it would require a pass through to know
what all the possible categorical values are first. Also, if I
understand correctly this would also be quite slow.
What I'm really looking for is suggestions about the best or idiomatic
"haskell" way to cumulatively build these results as it goes along, in
a reasonable amount of time. For instance, my knowledge of updatable
data structures in haskell isn't very good. Is there something obvious
or well-known that I'm overlooking? In general, I'm trying to move my
haskell knowledge from toy problem level to being useful for real
world problems.
Ian
On Tue, Jun 5, 2012 at 7:54 AM, Stephen Tetley <[email protected]> wrote:
> It's likely the while loop can be replaced with a fold which has a Map
> (or IntMap if the key is an Int) as the accumulator. For a large Map
> you will need to pay attention to inserting elements strictly...
>
> The more complicated bit is without any extra knowledge about the
> implementation of the "database" `get_next_from_db()` is currently
> magic.
------------------------------
Message: 6
Date: Tue, 5 Jun 2012 15:23:42 -0500
From: Gregory Guthrie <[email protected]>
Subject: [Haskell-beginners] graphics.gloss errors
To: "[email protected]" <[email protected]>
Message-ID:
<[email protected]>
Content-Type: text/plain; charset="us-ascii"
Why do I get this error, when trying the first line example from:
http://hackage.haskell.org/packages/archive/gloss/1.0.0.0/doc/html/Graphics-Gloss.html
>ghci
GHCi, version 7.4.1: http://www.haskell.org/ghc/ :? for help Loading package
ghc-prim ... linking ... done.
Loading package integer-gmp ... linking ... done.
Loading package base ... linking ... done.
Prelude> :m Graphics.Gloss
Prelude Graphics.Gloss> let main = displayInWindow "My Window" (200, 200) (10,
10) white (Circle 80)
<interactive>:4:12: Not in scope: `displayInWindow'
Also I notice that cabal shows the current version of GLFW as GLFW-0.5.0.1, But
their web-site shows it as 2.7.5 - big difference?!
I have:
Gloss 1.7.4.1
GLUT 2.3.0.0
GLFW 0.5.0.1
Opengl 2.5.0.0
------------------------------
Message: 7
Date: Tue, 5 Jun 2012 15:23:15 -0500
From: Gregory Guthrie <[email protected]>
Subject: [Haskell-beginners] wxHaskell install errors
To: "[email protected]" <[email protected]>
Message-ID:
<[email protected]>
Content-Type: text/plain; charset="us-ascii"
I had all of the wxHaskell programs installed and working fine, but now with an
update to Haskell Platform and thus also ghci, I need to upgrade all of them
(the previously working programs fail) requiring an update for the underlying
wxWidgets from 2.8 to 2.9 versions.
I downloaded and built wxWidgets 2.9, since the current wxHaskell libraries
require it (I couldn't find any existing binaries online).
The compile seemed to be successful, but when I then try to use it from the
wxHaskell programs I get an error that I don't know how to fix:
C: >cabal install wx
Resolving dependencies...
[1 of 1] Compiling Main ( C:\Users\guthrie\AppData\Local\Temp\wxc-0.
90.0.3-7324\wxc-0.90.0.3\Setup.hs, C:\Users\guthrie\AppData\Local\Temp\wxc-0.90.
0.3-7324\wxc-0.90.0.3\dist\setup\Main.o ) Linking
C:\Users\guthrie\AppData\Local\Temp\wxc-0.90.0.3-7324\wxc-0.90.0.3\dist\
setup\setup.exe ...
Configuring wxc-0.90.0.3...
Configuring wxc to build against wxWidgets 2.9
setup.exe: Missing dependencies on foreign libraries:
* Missing C libraries: wxmsw29ud_all, wxtiffd, wxjpegd, wxpngd,
wxzlibd,wxregexud, wxexpatd, wxregexud This problem can usually be solved by
installing the system packages that provide these libraries (you may need the
"-dev" versions). If the libraries are already installed but in a non-standard
location then you can use the flags --extra-include-dirs= and --extra-lib-dirs=
to specify where they are.
cabal: Error: some packages failed to install:
wx-0.90.0.1 depends on wxc-0.90.0.3 which failed to install.
wxc-0.90.0.3 failed during the configure step. The exception was: ExitFailure 1
wxcore-0.90.0.1 depends on wxc-0.90.0.3 which failed to install.
------------------------------
Message: 8
Date: Tue, 5 Jun 2012 16:33:30 -0400
From: Brent Yorgey <[email protected]>
Subject: Re: [Haskell-beginners] graphics.gloss errors
To: [email protected]
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
On Tue, Jun 05, 2012 at 03:23:42PM -0500, Gregory Guthrie wrote:
>
> Also I notice that cabal shows the current version of GLFW as
> GLFW-0.5.0.1, But their web-site shows it as 2.7.5 - big
> difference?!
I don't know the answer to the question about gloss. But re: the
above, note that the version of the GLFW library itself and the
version of the Haskell package called GLFW (which provides bindings to
the GLFW library) are completely different things. 0.5.0.1 appears to
be the latest version of the Haskell package.
-Brent
------------------------------
Message: 9
Date: Tue, 5 Jun 2012 15:36:36 -0500
From: Gregory Guthrie <[email protected]>
Subject: Re: [Haskell-beginners] graphics.gloss errors
To: "[email protected]" <[email protected]>
Message-ID:
<[email protected]>
Content-Type: text/plain; charset="us-ascii"
Apparently I needed to upgrade my examples, the API seems to have changed
across versions.
-------------------------------------------
> Subject: graphics.gloss errors
>
> Why do I get this error, when trying the first line example from:
>
> http://hackage.haskell.org/packages/archive/gloss/1.0.0.0/doc/html/Graphics-Gloss.html
>
> >ghci
> GHCi, version 7.4.1: http://www.haskell.org/ghc/ :? for help Loading package
> ghc-prim ...
> linking ... done.
> Loading package integer-gmp ... linking ... done.
> Loading package base ... linking ... done.
> Prelude> :m Graphics.Gloss
> Prelude Graphics.Gloss> let main = displayInWindow "My Window" (200, 200)
> (10, 10) white
> (Circle 80)
> <interactive>:4:12: Not in scope: `displayInWindow'
>
> Also I notice that cabal shows the current version of GLFW as GLFW-0.5.0.1,
> But their
> web-site shows it as 2.7.5 - big difference?!
>
> I have:
> Gloss 1.7.4.1
> GLUT 2.3.0.0
> GLFW 0.5.0.1
> Opengl 2.5.0.0
------------------------------
Message: 10
Date: Tue, 5 Jun 2012 21:48:38 +0100
From: Stephen Tetley <[email protected]>
Subject: Re: [Haskell-beginners] histogram over large data
To: Ian Knopke <[email protected]>
Cc: [email protected]
Message-ID:
<cab2tprbuezuuouj96d2tsxntfvgrp7973tacrudgcy9m849...@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Hi Ian
You really don't want to use a sort as it needs the whole list in
memory (or length as that counts its way through a list...).
If your data is integers and sparse - IntMap (provided you don't
overflow Int size) or the HashMaps suggested by Radoslaw should work.
If the data isn't sparse there are mutable arrays or (probably)
Data.Vector - I haven't got round to using the latter yet so can't
really comment on its suitability with respect to mutability.
If your data isn't integers you need to work out how to make a good
key from it so you can map key to count.
You can "fold" a Map or mutable array through a traversal of the data
- lists provide the usual fold (foldr) but if you have data stored in
a database there should be libraries for folding with an "iteraratee"
- again I haven't got round to using any of the iteratee packages yet
so couldn't comment on which one is the most appropriate.
Best wishes
Stephen
------------------------------
_______________________________________________
Beginners mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/beginners
End of Beginners Digest, Vol 48, Issue 5
****************************************
