I am reading a csv file containing 1 column with dates, 1 column with times and 120 columns of integers.
The file may contain many rows (could be over 100,000).

The data points are converted to color indexes between 0 and 1 and I plot them to a heatmap.

My problem is with displaying the times column as Y-axis labels.

Whatever I tried, the labels are displayed in a non-readable view.
Full view – labels are un-redable

If I zoom in to see more color details, the labels are more readable.
Zoom in – labels are readable

I would like to have the labels readable in any view, even when in full view.
It is acceptable that some of the labels will not be displayed in the zoom-out views and will only be displayed when zooming in.

I tried to create y ticks like below, where ‘data’ is a pandas dataframe and case_times is a dataframe with datetime objects.

import numpy as np
import pandas as pd
from datetime import datetime

data = pd.read_csv(csvfile, delimiter=r',|s||', header=None)
data = np.split(dsa_data, [1, 2, 62, 122], axis=1)
case_times = dsa_data[1]
case_times[1] = pd.to_datetime(case_times[1], format='%H:%M:%S').dt.time
case_times = np.squeeze(case_times)

fig, axs = plt.subplots(1,2)
fig.tight_layout()
im1 = axs[0].imshow(data, cmap='DSAColorMap')
    
    
axs[0].set_yticks(ticks=np.arange(1, len(case_times)+1, 1), labels=case_times)
axs[0].set_autoscale_on(True)
plt.show()

I then tried to start from a column of strings instead of the dataframe, but I get the same results:

with open(fileName, newline='') as csvfile:
    # Advance rows 3 time to skip headers and extract case start date 
    # without the need to iterate over the file
    row = next(csvfile)
    row = next(csvfile)
    row = next(csvfile) 
    text_to_split = re.split(r',|s||', row)
    case_start_date = text_to_split[0]
    # Remove blank items created after split by '|' delimiter
    text_to_split = text_to_split[:len(text_to_split)-3]
    #Create a list of time stamps to be later used as x axis for the plot
    case_times.append(text_to_split.pop(1)) 
    # Create the first data row
    data.append(text_to_split[1:61])
    spamreader = csv.reader(csvfile, delimiter='|')
    # Iterate over all remaining rows
    # to create the data rows
    for row in spamreader:
            text_to_split = ' '.join(row) # Create a list from all elements of data in a row
            text_to_split = re.split(r',|s', text_to_split) # split row by delimiters
            text_to_split.pop(len(text_to_split)-1) # Remove last element of the row which is blank
            text_to_split.pop(0) # Remove date from row
            #Create a list of time stamps to be later used as axis labels for the plot
            case_times.append(text_to_split.pop(0)) 
            # Add data row to each hemisphere data list
            data.append(text_to_split[:60])

converted_case_times = [] # transform from string to datetime
for i in range(len(case_times)):
    converted_case_times.append(datetime.strptime(case_times[i], '%H:%M:%S'))
    
fig, axs = plt.subplots(1,2)
fig.tight_layout()
im1 = axs[0].imshow(dsa_data_left, cmap='DSAColorMap')
axs[0].set_yticks(ticks=range(1, len(converted_case_times)+1), labels=converted_case_times)
plt.show()

I also tried the following formatting change to overcome the problem, but it also didn’t work:

from matplotlib.dates import DateFormatter

axs[0].yaxis.set_major_formatter(DateFormatter("%H:%M:%S"))
    plt.setp(axs[0].get_yticklabels(), rotation=45, ha="right",
         rotation_mode="anchor")

I assume the problem lies in the creation of the ticks (as a fixed range) however I do not know how to resolve this and I could not find an explanation or example that works in my case.

EDIT: Following the response I was given, I added this to my code:

    def generate_yticks(ax, case_times, ticks):
        ylims = ax.get_ylim()
        ticks_in_window = ticks[ticks>ylims[1]][ticks<ylims[0]]
        shown_ticks = 
        ticks_in_window[np.arange(10)*int(len(ticks_in_window)/10)]
        shown_labels = case_times[shown_ticks-1]
        ax.set_yticks(ticks=shown_ticks, labels=shown_labels)
        print(ylims)

   ticks = np.arange(1, len(case_times)+1)
   axs[0].callbacks.connect('ylim_changed', generate_yticks(axs[0], 
         case_times, ticks))

This seems to call the function only one time, meaning it is not called again when I zoom in.
I tried changing the ‘ylim_changed’ to another event, but I get an error of unsupported event.

If the issue is the method of using generate_ticks, i.e. that is has more than one parameter, I don’t know how to pass the ticks and case_times into the function, since it is a function called by another function.
I prefer not to make ticks a global variable, and I definitely cannot make case_times a global since it is generated within the code.

I tried to nest the generate_ticks function within the calling function, but that triggered the “boolean index did not match indexed array along axis 0; size of axis is 268 but size of corresponding boolean axis is 503” error again.