Matplotlib: show labels for minor ticks also

Question

In matplotlib, when I use a log scale on one axis, it might happen that that axis will have no major ticks, only minor

Answer 1

I've tried many ways to get minor ticks working properly in log plots. If you are fine with showing the log of the value of the tick you can use matplotlib.ticker.LogFormatterExponent. I remember trying matplotlib.ticker.LogFormatter but I didn't like it much: if I remember well it puts everything in base^exp (also 0.1, 0, 1). In both cases (as well as all the other matplotlib.ticker.LogFormatter*) you have to set labelOnlyBase=False to get minor ticks.

I ended up creating a custom function and use matplotlib.ticker.FuncFormatter. My approach assumes that the ticks are at integer values and that you want a base 10 log.

from matplotlib import ticker
import numpy as np

def ticks_format(value, index):
    """
    get the value and returns the value as:
       integer: [0,99]
       1 digit float: [0.1, 0.99]
       n*10^m: otherwise
    To have all the number of the same size they are all returned as latex strings
    """
    exp = np.floor(np.log10(value))
    base = value/10**exp
    if exp == 0 or exp == 1:   
        return '${0:d}$'.format(int(value))
    if exp == -1:
        return '${0:.1f}$'.format(value)
    else:
        return '${0:d}\\times10^{{{1:d}}}$'.format(int(base), int(exp))

subs = [1.0, 2.0, 3.0, 6.0]  # ticks to show per decade
ax.xaxis.set_minor_locator(ticker.LogLocator(subs=subs)) #set the ticks position
ax.xaxis.set_major_formatter(ticker.NullFormatter())   # remove the major ticks
ax.xaxis.set_minor_formatter(ticker.FuncFormatter(ticks_format))  #add the custom ticks
#same for ax.yaxis

If you don't remove the major ticks and use subs = [2.0, 3.0, 6.0] the font size of the major and minor ticks is different (this might be cause by using text.usetex:False in my matplotlibrc)

Answer 2

I think it is worth mentioning the option "minor_thresholds" introduced in matplotlib version 2.0 (docs link). It's a parameter in the form of a pair (subset, all) of the class LogFormatter that allows you to specify when a (fixed) subset of minor ticklabels should be shown and when all minor ticklabels should be shown (explanation of what this means is at the bottom).

In the following code I show the effect by using the same parameter values ((2, 0.4) in this case) but changing the limits of the x-axis:

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

fig, axes = plt.subplots(4, figsize=(12, 24))

dt = 0.01
t = np.arange(dt, 20.0, dt)

# first plot doesn't use a formatter
axes[0].semilogx(t, np.exp(-t / 5.0))
axes[0].set_xlim([0, 25])
axes[0].grid()

xlims = [[0, 25], [0.2, 8], [0.6, 0.9]]

for ax, xlim in zip(axes[1:], xlims):
    ax.semilogx(t, np.exp(-t / 5.0))
    formatter = LogFormatter(labelOnlyBase=False, minor_thresholds=(2, 0.4))
    ax.get_xaxis().set_minor_formatter(formatter)
    ax.set_xlim(xlim)
    ax.grid()

plt.show()

This results in the following plot:

You see that in the second row the ticklabels are the same as in the first row where we didn't use a formatter. This is because the log of the view-range is more than 2 (the first value of the parameter pair) or to put it differently, the view spans a range that is bigger than a range between two major ticklabels. In the third row the view is smaller than 2 but bigger than 0.4 (the second value of the parameter pair), so we see a subset of the minor ticklabels. Finally, in the last row the view spans less than 0.4 of the space between two major ticklabels, so all minor ticklabels are shown.

Answer 3

You can use set_minor_tickformatter on the corresponding axis:

from matplotlib import pyplot as plt
from matplotlib.ticker import FormatStrFormatter

axes = plt.subplot(111)
axes.loglog([3,4,7], [2,3,4])
axes.xaxis.set_minor_formatter(FormatStrFormatter("%.2f"))
plt.xlim(1.8, 9.2)
plt.show()