Description
Bug report
Bug summary
I've stumbled across something that can arise when point sizes are excessively large, so the only few lines that matter are
ax.scatter(xcoords, ycoords, s = pointsizes, **kwargs)
plt.savefig("figname.pdf")
The bug arose when an element of my pointsizes
array was unintentionally over 1e35
. This would save a PNG figure, but caused it to take excessively long in saving a PDF that I thought the program had stalled. The point sizes were based on the median value of a set of masses, some of which were very near zero. When I filtered the data more strictly to remove these outliers, the point sizes were much more reasonable and the figure was produced as expected - that is, once the pointsizes
array didn't have these excessively large values, the runtime was normal.
It may be useful to have a prescription for determining maximum allowed point sizes, or have a RuntimeWarning raised when point sizes are excessively large, if this isn't in the docs already.
Matplotlib version
- Operating system: Mac OS X Catalina
- Matplotlib version: 3.2.2
- Matplotlib backend (
print(matplotlib.get_backend())
): TkAgg - Python version: 3.7.4
- Jupyter version (if applicable): N/A
- Other libraries: N/A
Matplotlib installed via pip