Matplotlib - logarithmic scale, but require non-logarithmic labels

Question

How can I stop the y-axis displaying a logarithmic notation label on the y-axis?

I'm happy with the logarithmic scale, but want to display the absolute values, e.g. [500, 1500, 4500, 11000, 110000] on the Y-axis. I don't want to explicitly label each tick as the labels may change in the future (I've tried out the different formatters but haven't successfully gotten them to work). Sample code below.

Thanks,

-collern2

import matplotlib.pyplot as plt
import numpy as np

a = np.array([500, 1500, 4500, 11000, 110000])
b = np.array([10, 20, 30, 40, 50])

fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.set_yscale('log')

plt.plot(b, a)
plt.grid(True)
plt.show()

You may want to ask specific questions like this on matplotlib's user mailing list. — Thomas K, Jun 21 '11 at 20:07
I'm confused as I'm not sure what a "logarithmic notation label" is. Do you want to change the format of the labels (from 10^3 to 1000) or do you want to add ticks (or replace the current ticks by ticks) at the positions in a? — DSM, Jun 21 '11 at 20:32
The former - change the format of the labels (from 10^3 to 1000) — user809167, Jun 21 '11 at 20:34
10^3 is not logarithmic notation, it is scientific notation. — NeutronStar, Nov 11 '14 at 22:40

score 33 · Answer 1 · edited Aug 10 '18 at 23:40

33

If I understand correctly,

ax.set_yscale('log')

any of

ax.yaxis.set_major_formatter(matplotlib.ticker.ScalarFormatter())
ax.yaxis.set_major_formatter(matplotlib.ticker.FormatStrFormatter('%d'))
ax.yaxis.set_major_formatter(matplotlib.ticker.FuncFormatter(lambda x, pos: str(int(round(x)))))

should work. '%d' will have problems if the tick labels locations wind up being at places like 4.99, but you get the idea.

Note that you may need to do the same with the minor formatter, set_minor_formatter, depending on the limits of the axes.

edited Aug 10 '18 at 23:40

ImportanceOfBeingErnest

321,279
53
665
712

answered Jun 21 '11 at 20:42

DSM

342,061
65
592
494

+1 for also giving an example based on a `lambda`. This pretty much opens the door for every possible value-to-string mapping. Imho this answer should be accepted. – bluenote10 Feb 07 '14 at 19:01

score 3 · Answer 2 · answered Nov 04 '15 at 03:02

Use ticker.FormatStrFormatter

import matplotlib.pyplot as plt
import numpy as np
from matplotlib import ticker

a = np.array([500, 1500, 4500, 11000, 110000])
b = np.array([10, 20, 30, 40, 50])

fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.set_yscale('symlog')

ax.yaxis.set_major_formatter(ticker.FormatStrFormatter("%d"))

plt.plot(b, a)
plt.grid(True)

plt.show()

Matplotlib - logarithmic scale, but require non-logarithmic labels

2 Answers2

Linked

Related