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)