Repository navigation
OverflowError: In draw_path: Exceeded cell block limit #5907
Description
I am using matplotlib 1.5.1 on MacOSX to plot large datasets (roughly 2 million points) in a Jupyter notebook. When generating the plot using...
fig, ax = plt.subplots(1, 1)
panel_data.Gpi2.price.plot(ax=ax, logy=True)...I get the following error...
<matplotlib.axes._subplots.AxesSubplot at 0x13e9b4860>
Error in callback <function install_repl_displayhook.<locals>.post_execute at 0x10b6cc950> (for post_execute):
---------------------------------------------------------------------------
OverflowError Traceback (most recent call last)
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/pyplot.py in post_execute()
145 def post_execute():
146 if matplotlib.is_interactive():
--> 147 draw_all()
148
149 # IPython >= 2
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/_pylab_helpers.py in draw_all(cls, force)
148 for f_mgr in cls.get_all_fig_managers():
149 if force or f_mgr.canvas.figure.stale:
--> 150 f_mgr.canvas.draw_idle()
151
152 atexit.register(Gcf.destroy_all)
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/backend_bases.py in draw_idle(self, *args, **kwargs)
2024 if not self._is_idle_drawing:
2025 with self._idle_draw_cntx():
-> 2026 self.draw(*args, **kwargs)
2027
2028 def draw_cursor(self, event):
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/backends/backend_agg.py in draw(self)
472
473 try:
--> 474 self.figure.draw(self.renderer)
475 finally:
476 RendererAgg.lock.release()
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
59 def draw_wrapper(artist, renderer, *args, **kwargs):
60 before(artist, renderer)
---> 61 draw(artist, renderer, *args, **kwargs)
62 after(artist, renderer)
63
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/figure.py in draw(self, renderer)
1157 dsu.sort(key=itemgetter(0))
1158 for zorder, a, func, args in dsu:
-> 1159 func(*args)
1160
1161 renderer.close_group('figure')
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
59 def draw_wrapper(artist, renderer, *args, **kwargs):
60 before(artist, renderer)
---> 61 draw(artist, renderer, *args, **kwargs)
62 after(artist, renderer)
63
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/axes/_base.py in draw(self, renderer, inframe)
2322
2323 for zorder, a in dsu:
-> 2324 a.draw(renderer)
2325
2326 renderer.close_group('axes')
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
59 def draw_wrapper(artist, renderer, *args, **kwargs):
60 before(artist, renderer)
---> 61 draw(artist, renderer, *args, **kwargs)
62 after(artist, renderer)
63
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/lines.py in draw(self, renderer)
756 gc.set_sketch_params(*self.get_sketch_params())
757
--> 758 drawFunc(renderer, gc, tpath, affine.frozen())
759 gc.restore()
760
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/lines.py in _draw_lines(self, renderer, gc, path, trans)
1168
1169 def _draw_lines(self, renderer, gc, path, trans):
-> 1170 self._lineFunc(renderer, gc, path, trans)
1171
1172 def _draw_steps_pre(self, renderer, gc, path, trans):
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/lines.py in _draw_solid(self, renderer, gc, path, trans)
1193 def _draw_solid(self, renderer, gc, path, trans):
1194 gc.set_linestyle('solid')
-> 1195 renderer.draw_path(gc, path, trans)
1196
1197 def _draw_dashed(self, renderer, gc, path, trans):
/Users/drpugh/anaconda/lib/python3.4/site-packages/matplotlib/backends/backend_agg.py in draw_path(self, gc, path, transform, rgbFace)
164 self._renderer.draw_path(gc, p, transform, rgbFace)
165 else:
--> 166 self._renderer.draw_path(gc, path, transform, rgbFace)
167
168 def draw_mathtext(self, gc, x, y, s, prop, angle):
OverflowError: In draw_path: Exceeded cell block limit...oddly (at least to me) adding a plt.show() statement solved the problem...
fig, ax = plt.subplots(1, 1)
panel_data.Gpi2.price.plot(ax=ax, logy=True)
plt.show()...yields no error.
Activity
mdboom commented on Jan 25, 2016
plt.show() is probably triggering the macosx backend to render the figure, which does not have this limitation. Saving PNG files (and displaying images in %matplotlib inline mode in the Jupyter notebook) uses the agg backend, which does unfortunately have a hard-coded limit on the number of points in a line. You could try breaking your line up into pieces (we don't do this internally because it's difficult to do it in the general case), or consider other ways to reduce your data (seeing 2 million discrete points will be very difficult anyway).
tacaswell commented on Jan 25, 2016
I assume that this plot is super jagged? My guess is that there is something funny going on with when path simplification comes in or not. Which backend are you using?
davidrpugh commented on Jan 25, 2016
@tacaswell Using the plt.get_backend() command returns...
'module://ipykernel.pylab.backend_inline'
...so inline plotting.
wavexx commented on Feb 2, 2016
I encountered the same problem today, on linux (python 3, matplotlib 1.5.1~rc1 from debian).
I'm plotting large time series (101k points). The line plot is dense enough that it ends up looking like a filled area between 2 jagged curves, which is exactly the desired effect, without having to generate the area myself.
I can show() the plot just fine using the Qt4Agg backend, but cannot savefig().
wavexx commented on Feb 2, 2016
Or, should I clarify, I can show() it as along as the window size is small enough. I discovered now that I cannot get it big enough to screenshot either...
tacaswell commented on Feb 3, 2016
Try setting mpl.rcParams['agg.path.chunksize'] = 10000 or so.
wavexx commented on Feb 4, 2016
On 03/02/16 05:35, Thomas A Caswell wrote:
That does the trick (both for Qt4Agg and plain Agg).
Maybe that should go as a commented note in the default matplotlibrc.
anntzer commented on Sep 11, 2017
This was documented in #8176.
abatoolyn commented on Nov 27, 2018
Im using
votes.plot(x='Lat', y='Lon')
to plot points as a map
What can I do to overcome the exceeded cell block limit?