why does matplotlib return an object instead of plotting it?

Question

I'm running this in the current version of Jupyter.

One of my cells contains the following.

import matplotlib.pyplot as plt

def main():  
    a_list = list(range(20))
    plt.plot(a_list)  
    plt.show()

main()

When I run kernel > Restart & Run All the only output I get is a plot figure: <matplotlib.figure.Figure at 0x1bd35aa7a20> .

When I run the cell directly with the Run button, I get the expected plot.

When I move main() to its own cell and then run kernel > Restart & Run All I get the expected plot.

I would appreciate help in understanding what's going on.

If you place `%matplotlib inline` at the beginning of your notebook and then run `kernel > Restart & Run All`, do you get the expected plot? (If so, then [here is a way](https://stackoverflow.com/q/21176731/190597) to automatically run `%matplotlib inline` whenever IPython notebook is started.) — unutbu, May 28 '18 at 02:00
Yes, adding `%matplotlib inline` lets the initial example work as expected. But still, what's going on? Why is that necessary (sometimes)? And how would someone know to do that? — RussAbbott, May 28 '18 at 02:05
I'm not quite sure, but I believe the issue here is that there are a number of possible [matplotib backends](https://matplotlib.org/tutorials/introductory/usage.html#backends) available and IPython/Jupyter is not picking the right one / has not been configured to pick the one you desire. See [this question](https://stackoverflow.com/q/43027980/190597) which has links to the docs. — unutbu, May 28 '18 at 02:17

0 Answers0