Label and tick marks disappear when using log scale in matplotlib python

Question

I need the y-axis of my graph to be on a log scale. When I do so, however, the y-axis' label, tick marks, and title disappear.

plt.figure(2)
plt.semilogy(data2[0, :, 0], sli)
plt.xlabel('n-value')
plt.ylabel('Intensity')
plt.title('Intensity vs. n-shell')
plt.show()

The sli values range from 1.0e-21 to 1.0e-8

I'm not cool enough to post images

When I zoom in far enough though, the label and title actually return but not the tick marks. Don't know if that matters, but thought I'd include it.

!still not cool enough

Thanks

Edit: As it turns out, the code works fine, just not on my mac laptop. I tested the code on a friends computer running ubuntu and it worked perfectly. So, I guess my log scales don't like macs. Still, anybody have any suggestions?

"Update" for @ImportanceOfBeingErnest

Nothing has changed.

Graph produced from my code run with updated mplib

Graph produced from @Engineero 's code run with updated mplib

"log scales don't like macs" - that's an interesting theory, but I suspect that you rather simply have a buggy version of matplotlib installed. You may update matplotlib and see if that fixes it. — ImportanceOfBeingErnest, Jun 13 '18 at 01:15
Under your recommendation, I reinstalled matplotlib and checked the version i'm using (2.2.2). Still no luck :/ — pierre700, Jun 13 '18 at 01:50
Can you update your question with a [mcve], i.e. a code one can run and the images you obtain from that updated version? — ImportanceOfBeingErnest, Jun 13 '18 at 01:59
Added a section under the bold for you. Nothing has changed as far as I can tell. — pierre700, Jun 13 '18 at 02:23
Yes, seems like you aren't using the matplotlib version that you installed but still the old one. — ImportanceOfBeingErnest, Jun 13 '18 at 09:41
Just tried updating again. This, among other things popped up: `Requirement already up-to-date: matplotlib in /Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages` Still same result — pierre700, Jun 13 '18 at 16:10
Sure, you have probably installed that correctly, but you are not *using* it when running your code. Add `import matplotlib; print(matplotlib.__version__)` inside the code that produces those images for you. — ImportanceOfBeingErnest, Jun 13 '18 at 16:29
Do you know how I can start using the correct version of matplotlib? — pierre700, Jun 14 '18 at 21:22
You need to use the correct python version, i.e the python version for which you installed matplotlib. — ImportanceOfBeingErnest, Jun 14 '18 at 22:10
Ahhhhh!!! Yup, apparently I was running python 2.7 but I thought I was running 3. I changed my build system to run python 3 and it works now! Thanks a ton!!! — pierre700, Jun 14 '18 at 23:01

score 1 · Answer 1 · answered Jun 12 '18 at 21:21

1

Try using axis.set_yscale with axis.tick_params. Something like:

fig = plt.figure(2)
axis = fig.add_subplot(111)
axis.plot(data2[0, :, 0], sli)
axis.set_yscale('log', nonposy='clip')
axis.tick_params(axis='y', which='minor', colors='black')
axis.set_xlabel('n-value')
axis.set_ylabel('Intensity')
axis.set_title('Intensity vs. n-shell')
plt.show()

Basically use the axis API. This is the only way I was ever able to get minor log-scale tick marks to work for me the way that I wanted...

answered Jun 12 '18 at 21:21

Engineero

12,340
5
53
75

I plugged your code in and got the exact same thing as my code. Still getting the "error." I'm guessing it has something to do with my mac since the code actually works fine on my friend's linux. Thanks for your contribution though! – pierre700 Jun 12 '18 at 21:58
@pierre700 Try adding `matplotlib.use('TkAgg')` before you plot. If that doesn't work, try just 'Agg', maybe 'GTK'. – Engineero Jun 12 '18 at 22:02
No luck. Tried all of them and they made what would be the graph just appear as a white window. – pierre700 Jun 12 '18 at 22:18
@pierre700 try the different backends listed [here](https://stackoverflow.com/a/43030607/3670871). If none of those work then I'm out of guesses. – Engineero Jun 12 '18 at 22:23

Label and tick marks disappear when using log scale in matplotlib python

1 Answers1