matplotlib-users

[Matplotlib-users] problem plotting large datasets

[Antonio G. <Ant...@ki...>]

Dear all,

I have problems displaying plots for large datasets. Briefly, when attemp=
ting to=20
plot a large number of data points these are not displayed, or only parti=
ally=20
displayed. The simplest way for reproducing my problem is:

   from pylab import *
   datalen =3D 100000

   data =3D rand(datalen)
   plot(data)
   show()

Here, if datalen is 10000, then the graph looks as expected. However, if =
datalen=20
is increased, say, to 100000, then the resulting plot is incomplete (as s=
hown in=20
snapshot1). Moreover, when trying to zoom, the axes display the new value=
s=20
reflecting the zoomed scale, but the data are not actually zoomed (eg=20
snapshot2). I have no clue as to what can be wrong, since no error is ret=
urned.=20
Also, no error results from running the example above with --verbose-help=
ful:

   antonio@linux:~/python> python test.py --verbose-helpful
   matplotlib data path /usr/lib/python2.4/site-packages/matplotlib/mpl-d=
ata
   $HOME=3D/home/antonio
   CONFIGDIR=3D/home/antonio/.matplotlib
   loaded rc file /home/antonio/.matplotlib/matplotlibrc
   matplotlib version 0.88svn
   verbose.level helpful
   interactive is False
   platform is linux2
   numerix numpy 0.9.6.2206
   font search path ['/usr/lib/python2.4/site-packages/matplotlib/mpl-dat=
a']
   loaded ttfcache file /home/antonio/.matplotlib/ttffont.cache
   backend TkAgg version 8.4

In a real example of what I'm trying to do: I read a fragment of data fro=
m a=20
large file, then plot such fragment of data; convolve it and plot the res=
ult (as=20
in snapshot3); interactively make some decisions according to such plot, =
and=20
then move to the next fragment of my large file to repeat the process. Ho=
wever,=20
whenever I attempt to plot large chunks of data, the plot is not displaye=
d (eg=20
snapshot4). As mentioned, no error is returned, and thus I have no clue a=
s to=20
how to fix this problem.

(For the record, in snapshot3 the length of data plotted is 671334 [top] =
and=20
132829 [bottom]. In snapshot4 it is 1162554 [top] and 231073 [bottom].)

Any help will be much appreciated.

Antonio



--=20
Antonio Gonzalez, MD MSc PhD
Department of Neuroscience
Karolinska Institutet
Retzius v=C3=A4g. 8
171 77 Stockholm
Sweden

Re: [Matplotlib-users] problem plotting large datasets

[Darren D. <dd...@co...>]

Hi Antonio,

This is a blast from the past, it's a limitation of AGG:
http://sourceforge.net/mailarchive/message.php?msg_id=9175398

Darren


On Wednesday 22 March 2006 7:13 am, Antonio Gonzalez wrote:
> Dear all,
>
> I have problems displaying plots for large datasets. Briefly, when
> attempting to plot a large number of data points these are not displayed,
> or only partially displayed. The simplest way for reproducing my problem
> is:
>
>    from pylab import *
>    datalen = 100000
>
>    data = rand(datalen)
>    plot(data)
>    show()
>
> Here, if datalen is 10000, then the graph looks as expected. However, if
> datalen is increased, say, to 100000, then the resulting plot is incomplete
> (as shown in snapshot1). Moreover, when trying to zoom, the axes display
> the new values reflecting the zoomed scale, but the data are not actually
> zoomed (eg snapshot2). I have no clue as to what can be wrong, since no
> error is returned. Also, no error results from running the example above
> with --verbose-helpful:
>
>    antonio@linux:~/python> python test.py --verbose-helpful
>    matplotlib data path
> /usr/lib/python2.4/site-packages/matplotlib/mpl-data $HOME=/home/antonio
>    CONFIGDIR=/home/antonio/.matplotlib
>    loaded rc file /home/antonio/.matplotlib/matplotlibrc
>    matplotlib version 0.88svn
>    verbose.level helpful
>    interactive is False
>    platform is linux2
>    numerix numpy 0.9.6.2206
>    font search path
> ['/usr/lib/python2.4/site-packages/matplotlib/mpl-data'] loaded ttfcache
> file /home/antonio/.matplotlib/ttffont.cache
>    backend TkAgg version 8.4
>
> In a real example of what I'm trying to do: I read a fragment of data from
> a large file, then plot such fragment of data; convolve it and plot the
> result (as in snapshot3); interactively make some decisions according to
> such plot, and then move to the next fragment of my large file to repeat
> the process. However, whenever I attempt to plot large chunks of data, the
> plot is not displayed (eg snapshot4). As mentioned, no error is returned,
> and thus I have no clue as to how to fix this problem.
>
> (For the record, in snapshot3 the length of data plotted is 671334 [top]
> and 132829 [bottom]. In snapshot4 it is 1162554 [top] and 231073 [bottom].)
>
> Any help will be much appreciated.
>
> Antonio

-- 
Darren S. Dale, Ph.D.
dd...@co...

Re: [Matplotlib-users] problem plotting large datasets

[Antonio G. <Ant...@ki...>]

Darren Dale wrote:
> Hi Antonio,
>=20
> This is a blast from the past, it's a limitation of AGG:
> http://sourceforge.net/mailarchive/message.php?msg_id=3D9175398
>=20
> Darren

Thanks for replying, Darren. Actually, while searching for an answer to m=
y=20
problem a few days ago, I read the thread that you point out. It's just t=
hat,=20
since such thread is already a couple of years old, I was hoping someone =
had=20
come with a way of sorting the problem out.

Anyway, I'll cut my data down to some useful length. Thanks,

Antonio



--=20
Antonio Gonzalez, MD MSc PhD
Department of Neuroscience
Karolinska Institutet
Retzius v=C3=A4g. 8
171 77 Stockholm
Sweden