matplotlib

On Monday 23 June 2008 23:37:09 Adam Mercer wrote:
> Hi
>
> Just upgraded to matplotlib-0.98.1, and a code that worked with 0.98.0
> is now failing with the following error:

Please, include a brief standalone script that demonstrates the problem when 
reporting problems.

> Traceback (most recent call last):
>   File "./plot_workout.py", line 126, in <module>
>     time_plot = time_axes.plot_date(times_dates, times, 'bo-')
>   File "/opt/local/lib/python2.5/site-packages/matplotlib/axes.py",
> line 3081, in plot_date
>     self.xaxis_date(tz)
>   File "/opt/local/lib/python2.5/site-packages/matplotlib/axes.py",
> line 2015, in xaxis_date
>     locator.refresh()
>   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> line 540, in refresh
>     dmin, dmax = self.viewlim_to_dt()
>   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> line 436, in viewlim_to_dt
>     return num2date(vmin, self.tz), num2date(vmax, self.tz)
>   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> line 233, in num2date
>     if not cbook.iterable(x): return _from_ordinalf(x, tz)
>   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> line 156, in _from_ordinalf
>     dt = datetime.datetime.fromordinal(ix)
> ValueError: ordinal must be >= 1
>
> Any ideas?
>
> Cheers
>
> Adam
>
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://sourceforge.net/services/buy/index.php
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users@...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

I used to have such errors when I was trying to display empty vectors.
But since you're saying that the script work on previous version of
Matplotlib...
Sorry if it hasn't helped.
Johan

2008/6/24 Darren Dale <dsdale24@...>:

> On Monday 23 June 2008 23:37:09 Adam Mercer wrote:
> > Hi
> >
> > Just upgraded to matplotlib-0.98.1, and a code that worked with 0.98.0
> > is now failing with the following error:
>
> Please, include a brief standalone script that demonstrates the problem
> when
> reporting problems.
>
> > Traceback (most recent call last):
> >   File "./plot_workout.py", line 126, in <module>
> >     time_plot = time_axes.plot_date(times_dates, times, 'bo-')
> >   File "/opt/local/lib/python2.5/site-packages/matplotlib/axes.py",
> > line 3081, in plot_date
> >     self.xaxis_date(tz)
> >   File "/opt/local/lib/python2.5/site-packages/matplotlib/axes.py",
> > line 2015, in xaxis_date
> >     locator.refresh()
> >   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> > line 540, in refresh
> >     dmin, dmax = self.viewlim_to_dt()
> >   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> > line 436, in viewlim_to_dt
> >     return num2date(vmin, self.tz), num2date(vmax, self.tz)
> >   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> > line 233, in num2date
> >     if not cbook.iterable(x): return _from_ordinalf(x, tz)
> >   File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
> > line 156, in _from_ordinalf
> >     dt = datetime.datetime.fromordinal(ix)
> > ValueError: ordinal must be >= 1
> >
> > Any ideas?
> >
> > Cheers
> >
> > Adam
> >
> > -------------------------------------------------------------------------
> > Check out the new SourceForge.net Marketplace.
> > It's the best place to buy or sell services for
> > just about anything Open Source.
> > http://sourceforge.net/services/buy/index.php
> > _______________________________________________
> > Matplotlib-users mailing list
> > Matplotlib-users@...
> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>
>
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://sourceforge.net/services/buy/index.php
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users@...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

On Tue, Jun 24, 2008 at 6:43 AM, Johan Mazel <johan.mazel@...> wrote:
> I used to have such errors when I was trying to display empty vectors.
> But since you're saying that the script work on previous version of
> Matplotlib...

I recently made some changes to try and support a use-case where the
user has no data and sets ax.set_xdate and later plots some date data.

Eg, the following used to fail on the call to xaxis_date because the
default data/viewlim are 0..1

  import datetime
  import matplotlib.pyplot as plt
  dt = datetime.date
  x = dt(2002,1,1), dt(2003,1,1), dt(2004,1,1)
  fig = plt.figure()
  ax = fig.add_subplot(111)
  ax.xaxis_date()
  ax.plot(x, [1,2,3])
  fig.autofmt_xdate()
  plt.show()

so it is possible these edits are causing the change Adam is
describing (even though they were designed to fix it), but as Darren
notes we will need a standalone script since our examples are working
fine and it is hard to debug in the void.

JDH

On Tue, Jun 24, 2008 at 6:14 AM, Darren Dale <dsdale24@...> wrote:

> Please, include a brief standalone script that demonstrates the problem when
> reporting problems.

Sorry should have done that, I've attached an example script (and the
according data files) which exhibits the problem on 0.98.1, 0.98.0 has
no problems.

  File "./test.py", line 46, in <module>
    test_plot = test_axes.plot_date(test_dates, test, 'bo-')
  File "/opt/local/lib/python2.5/site-packages/matplotlib/axes.py",
line 3081, in plot_date
    self.xaxis_date(tz)
  File "/opt/local/lib/python2.5/site-packages/matplotlib/axes.py",
line 2015, in xaxis_date
    locator.refresh()
  File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
line 540, in refresh
    dmin, dmax = self.viewlim_to_dt()
  File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
line 436, in viewlim_to_dt
    return num2date(vmin, self.tz), num2date(vmax, self.tz)
  File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
line 233, in num2date
    if not cbook.iterable(x): return _from_ordinalf(x, tz)
  File "/opt/local/lib/python2.5/site-packages/matplotlib/dates.py",
line 156, in _from_ordinalf
    dt = datetime.datetime.fromordinal(ix)
ValueError: ordinal must be >= 1

What's strange is that if I comment out the plotting of the second
test data set then the plot is produced without error, even though the
reported error (when plotting both data sets) seems to have nothing to
do with the second data set.

Cheers

Adam

On Tue, Jun 24, 2008 at 8:13 AM, Adam Mercer <ramercer@...> wrote:

> What's strange is that if I comment out the plotting of the second
> test data set then the plot is produced without error, even though the
> reported error (when plotting both data sets) seems to have nothing to
> do with the second data set.

Hey Adam,

I think I have this fixed in svn -- also, I rewrote your example to
use csv2rec (which returns record arrays).  Thought you might be
interested:

import matplotlib.mlab as mlab
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import matplotlib.cbook as cbook
r1 = mlab.csv2rec('test1.dat', names='date,value', delimiter=' ',
                  converterd={'date':cbook.todate('%Y%m%d')})
r2 = mlab.csv2rec('test2.dat', names='date,value', delimiter=' ',
                  converterd={'date':cbook.todate('%Y%m%d')})

print 'dtype', r1.dtype
fig = plt.figure()
ax1 = fig.add_subplot(111)
ax2 = ax1.twinx()

# produce plot
line1, = ax1.plot(r1.date, r1.value, 'bo-')
line2,  = ax2.plot(r2.date, r2.value, 'bo-')

# set up axes
ax1.xaxis.set_major_locator(mdates.DayLocator(range(0, 31, 2)))
ax1.xaxis.set_minor_locator(mdates.DayLocator(range(0, 31, 1)))
ax1.xaxis.set_major_formatter(mdates.DateFormatter('%Y/%m/%d'))

plt.show()

> I think I have this fixed in svn -- also, I rewrote your example to
> use csv2rec (which returns record arrays).  Thought you might be
> interested:
>
> import matplotlib.mlab as mlab
> import matplotlib.pyplot as plt
> import matplotlib.dates as mdates
> import matplotlib.cbook as cbook
> r1 = mlab.csv2rec('test1.dat', names='date,value', delimiter=' ',
>                  converterd={'date':cbook.todate('%Y%m%d')})
> r2 = mlab.csv2rec('test2.dat', names='date,value', delimiter=' ',
>                  converterd={'date':cbook.todate('%Y%m%d')})
>
> print 'dtype', r1.dtype
> fig = plt.figure()
> ax1 = fig.add_subplot(111)
> ax2 = ax1.twinx()
>
> # produce plot
> line1, = ax1.plot(r1.date, r1.value, 'bo-')
> line2,  = ax2.plot(r2.date, r2.value, 'bo-')
>
> # set up axes
> ax1.xaxis.set_major_locator(mdates.DayLocator(range(0, 31, 2)))
> ax1.xaxis.set_minor_locator(mdates.DayLocator(range(0, 31, 1)))
> ax1.xaxis.set_major_formatter(mdates.DateFormatter('%Y/%m/%d'))
>
> plt.show()

Thanks John, although the scale on the x-axis isn't shown using the
above example.

Cheers

Adam

On Wed, Jun 25, 2008 at 8:44 AM, John Hunter <jdh2358@...> wrote:

> I think I have this fixed in svn

Just downloaded 0.98.2 and the code that failed in 0.98.1 works! Thanks!

Cheers

Adam