Format time axis to "2020-JAN-2021-JAN"

Community
1

I’m trying to find a way to format the datetime axis to display years as numbers and months as uppercase letters. Unfortunately, Python’s strptime function doesn’t provide uppercase month names (e.g. APR). I need to format it myself, but I haven’t been able to figure out how. Since I only want to show Apr, Jul, Oct. I used a MonthLocator in a ConciseDateFormatter. I have tried the following:

import pandas as pd
import numpy as np

import matplotlib.pyplot as plt
from matplotlib.dates import MonthLocator
from matplotlib.dates import ConciseDateFormatter
from matplotlib.ticker import FuncFormatter

x = pd.date_range('2016-01-01', '2021-12-31', freq='M')
y = np.random.rand(len(x))

fig, ax = plt.subplots(figsize=(10, 8))
locator = MonthLocator((1, 4, 7, 10))
ax.xaxis.set_major_locator(locator)
ax.xaxis.set_major_formatter(
    ConciseDateFormatter(
        locator, formats=['%Y', '%b', '%d', '%H:%M', '%H:%M', '%S.%f']
    )
)
ax.tick_params(axis='x', labelsize=8)
ax.plot(x, y)

3b6678fc-35e8-4120-be76-c107f05b19c5

But I want the axis to be

2016 - APR - JUL - OCT - 2017 - APR - JUL -OCT - 2018 -

I’ve tried a few ways, but none of them works. Once I use a user-defined FuncFormatter, it will no longer respect my ConciseDateFormatter. I’m wondering what the easiest way is for me to achieve my desired outcome. Should I simply write a formatter function that includes an if statement? If the year and month are both equal to 1, it should display the year number; otherwise, it should show the month name. If this is the case, how can I accomplish it?

Thanks in advance.

2

Have you tried using set_ticklabels? Something like

ax.tick_params(axis='x', labelsize=8)
ax.plot(x, y)
ax.xaxis.set_ticklabels([t.get_text().upper() 
                         for t in ax.get_xticklabels()
                         ]);

Matplotlib issues a UserWarning: set_ticklabels() should only be used with a fixed number of ticks, i.e. after set_ticks() or using a FixedLocator, so this may not work in all circumstances.

3

Thank you! Indeed, I believe this is a way to format the labels instead of resetting them manually.

1 Like
4
5

Another approach would be to create a ConciseDateFormatter subclass that overrides format_ticks. This approach may not work for a wide range of Matplotlib versions and could break if the ConciseDateFormatter code changes in a future update. This works in Matplotlib 3.7.3 and 3.8.0

import pandas as pd
...
from matplotlib.dates import ConciseDateFormatter
import calendar   # for month names in your locale

class UpperConciseDateFormatter(ConciseDateFormatter):
    def format_ticks(self, values):
        labels = super().format_ticks(values)
        if not self._usetex:
            labels =  [label.upper()
                       for label in labels]
        else:
            months = [calendar.month_abbr[m] for m in range(1, 13)]
            labels = [label.upper() if label in months else label
                      for label in labels]
        return labels

x = pd.date_range('2016-01-01', '2021-12-31', freq='M')
...
ax.xaxis.set_major_formatter(
    UpperConciseDateFormatter(
        locator, formats=['%Y', '%b', '%d', '%H:%M', '%H:%M', '%S.%f']
    )
)
ax.tick_params(axis='x', labelsize=8)
ax.plot(x, y)
1 Like