matplotlib bug in plot_dates?

Question

Just wanted to see if others thought that the following behaviour in matplotlib plot_date was buggy, or if it's just something I should put up with.

I have a multi-panel plot that I set up with sharex to facilitate zoom/pan on all axes, and I plot time series data in both panels. However, in the second panel, all of the data happens to be invalid (in this example I mask it).

from matplotlib.pyplot import figure,show
from datetime import datetime,timedelta
from numpy import sin,cos,linspace,pi,ma,array

fig=figure(figsize=(16,9))
ax1=fig.add_subplot(211)
ax2=fig.add_subplot(212,sharex=ax1)

# xdata is seconds
xdata=linspace(0,9999,10000)
tdata=array([datetime(2000,1,1)+timedelta(seconds=ss) for ss in xdata])
data1=ma.masked_array(sin(pi*xdata/300),mask=False)
data2=ma.masked_array(cos(pi*xdata/300),mask=True)

ax1.plot_date(tdata,data1,marker='',color='r')
ax2.plot_date(tdata,data2,marker='',color='b')
show()

I'd expect (prefer) it to just show up a blank panel, not to fail and give me a long unhelpful traceback. Is this the expected behaviour?

Notes:

This script also fails if I use ax.plot(...) instead of ax.plot_date(...)
It works fine (i.e. gives me an empty panel) if I just plot against xdata, not the datetime array tsdata (but I have to use ax1.set_xlim(xdata[0],xdata[-1]) to get a sensible domain displayed):

   ax1.plot(xdata,data1,marker='',color='r')
   ax2.plot(xdata,data2,marker='',color='b')
   ax1.set_xlim(xdata[0],xdata[-1])
   show()

I have just realised I can rescue the above plot by forcing the ax2 limits right before the show() command. I still think that the failure in the main example is inelegant:

   ax2.set_xlim(tdata[0],tdata[-1])
   show()

What do the experts think?

Thanks!

F.Y.I., This was on matplotlib 1.1.0, compiled from source on my PC.

Here is the traceback that I get:

Traceback (most recent call last):

 File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/backendsbackend_gtk.py", line 395, in expose_even
self._render_figure(self._pixmap, w, h)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/backends/backend_gtkagg.py", line 75, in _render_f
FigureCanvasAgg.draw(self)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/backends/backend_agg.py", line 401, in draw
self.figure.draw(self.renderer)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/artist.py", line 55, in draw_wrapper
draw(artist, renderer, *args, **kwargs)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/figure.py", line 884, in draw
func(*args)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/artist.py", line 55, in draw_wrapper
draw(artist, renderer, *args, **kwargs)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/axes.py", line 1983, in draw
a.draw(renderer)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/artist.py", line 55, in draw_wrapper
draw(artist, renderer, *args, **kwargs)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/axis.py", line 1036, in draw
ticks_to_draw = self._update_ticks(renderer)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/axis.py", line 926, in _update_ticks
tick_tups = [ t for t in self.iter_ticks()]
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/axis.py", line 873, in iter_ticks
majorLocs = self.major.locator()
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/dates.py", line 749, in __call__
self.refresh()
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/dates.py", line 758, in refresh
dmin, dmax = self.viewlim_to_dt()
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/dates.py", line 530, in viewlim_to_dt
return num2date(vmin, self.tz), num2date(vmax, self.tz)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/dates.py", line 289, in num2date
if not cbook.iterable(x): return _from_ordinalf(x, tz)
  File "/home/tchubb/local/lib/python2.7/site-packages/matplotlib/dates.py", line 203, in _from_ordinalf
dt = datetime.datetime.fromordinal(ix)
ValueError: ordinal must be >= 1

If you suspect this is an IPython bug, please report it at:
https://github.com/ipython/ipython/issues
or send an email to the mailing list at [email protected]

could you paste up the trace back you get so people can look at it with out having to run your code? What I get is in my answer, but it would be good to make sure I am getting the same error — tacaswell
– tacaswell, Commented Oct 7, 2013 at 1:17
What version of matplotlib are you using? I can't reproduce this in either 1.2.1 or 1.3.0 — ali_m
– ali_m, Commented Oct 7, 2013 at 1:21
Looks like it might be this bug. It ought to be fixed from v1.2 onwards. — ali_m
– ali_m, Commented Oct 7, 2013 at 1:25

tacaswell · Accepted Answer · 2013-10-07 01:47:49Z

Yes, I would call this a bug, or at least an oversight which looks like it has been fixed (via @ali_m https://github.com/matplotlib/matplotlib/issues/162).

I was getting this exception, but that was on a version of 1.3.x from June (my bad, thought I had a more current version on this computer). Current master does not have this problem and @ali_m reports that it also works on 1.2.1 and 1.3.0 so I suspect the fix is to upgrade you version of matplotlib.

What looks like is going on is that there isn't a check in the code that sorts out where to put the ticks has no check to make sure you have given in non-empty data.

The reason that setting the limits explicitly prevents the error is that the code never tries to figure out what the range of your empty data is.

Please create a github issue for this (and include the trace back in the issue).

In [7]: Traceback (most recent call last):
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/backends/backend_qt4.py", line 366, in idle_draw
    self.draw()
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/backends/backend_qt4agg.py", line 148, in draw
    FigureCanvasAgg.draw(self)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/backends/backend_agg.py", line 440, in draw
    self.figure.draw(self.renderer)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/artist.py", line 54, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/figure.py", line 1027, in draw
    func(*args)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/artist.py", line 54, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/axes.py", line 2088, in draw
    a.draw(renderer)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/artist.py", line 54, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/axis.py", line 1076, in draw
    ticks_to_draw = self._update_ticks(renderer)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/axis.py", line 938, in _update_ticks
    tick_tups = [t for t in self.iter_ticks()]
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/axis.py", line 882, in iter_ticks
    majorLocs = self.major.locator()
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/dates.py", line 785, in __call__
    self.refresh()
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/dates.py", line 794, in refresh
    dmin, dmax = self.viewlim_to_dt()
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/dates.py", line 560, in viewlim_to_dt
    return num2date(vmin, self.tz), num2date(vmax, self.tz)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/dates.py", line 305, in num2date
    return _from_ordinalf(x, tz)
  File "/home/tcaswell/local_installs/lib/python2.7/site-packages/matplotlib-1.3.x-py2.7-linux-x86_64.egg/matplotlib/dates.py", line 208, in _from_ordinalf
    dt = datetime.datetime.fromordinal(ix)
ValueError: ordinal must be >= 1

Thanks for your response. Interestingly a previous commenter was unable to reproduce the error on a more recent version of matplotlib than I used, but you have version 1.3, looks like you compiled from source (as did I).
Thanks! I'm happy to take that at face value. I'll get around to upgrading one day...

Collectives™ on Stack Overflow

matplotlib bug in plot_dates?

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related