如何减少Matplotlib对数图上的主要刻度间距

Question

I would like to change the major tick label spacing to intervals of 0.1 (base 0.1) and have minor ticks (without labels) at 0.01 spacing for the code below (output shown in 1 .

x = np.array([1, 5, 10, 32])
y = np.array([0.34, 0.27, 0.25, 0.21])
yerr = np.array([0.02, 0.019, 0.019, 0.016])

fig, ax = plt.subplots()
res = ax.errorbar(x, y, yerr=yerr, fmt='.', ls='')

ax.set_xscale('log')
ax.set_yscale('log')

I've tried to do

import matplotlib.ticker as mtick
ax.yaxis.set_major_locator(mtick.LogLocator(base=0.01))

however nothing changes, which makes me think the base can not be less than 10.

Answer 1

EDIT: OK, the reasoning behind this is wrong ( & I misundertood the question)

This would add equidistant space between the ticks, definitely not you were asking.

plt.grid("on")  # Just to see the spacing more clearly
ax.yaxis.set_major_locator(mtick.LogLocator(base=10**(1.0/1)))
ax.yaxis.set_minor_locator(mtick.LogLocator(base=10**(1.0/10)))

(asuming 10 is the base of the log scale you're using and the fraction is how many more (or less) lines you want.

I did not understand how the decimal values in the 'base' variable work when using logarithmic scale.

---

Well, since you're working in logarithmic values ... that 'spacing' shouldn't be 1 or less. (using 1 you get an error) if you want to reduce spacing, just use a value between less than 10 (but more than 1).

In fact, using the line:

plt.grid("on")  # Just to see the spacing more clearly
ax.yaxis.set_major_locator(mtick.LogLocator(base=10**(1/10)))

you get a tenth of a spacing than before.

It's all in how the vertical axes is computed when using the logarithmic scale.

Answer 2

SOLUTION: Thanks ImportanceOfBeingErnest for the advice. My solution is to add this:

import matplotlib.ticker as mtick
ax.yaxis.set_major_locator(mtick.LogLocator(base=10, subs=range(10)))
ax.yaxis.set_minor_locator(mtick.LogLocator(base=10, subs=range(100)))
plt.setp(ax.get_yminorticklabels(), visible=False);