Regarding subplot and sub-sub plot location in matplotlib-CodePudding

I am having some trouble understanding the subplot concept. As I understand subplots are organized (number of rows, numbers of column, and the plot number or how many places the plot is taking)

I have this code:

from pandas import util
import pandas.util.testing as testing
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
import matplotlib as mpl

df = util.testing.makeDataFrame()


with mpl.rc_context(rc={'font.family': 'serif', 'font.weight': 'light', 'font.size': 12}):
    fig = plt.figure(figsize= (12, 6))

    
    ax = fig.add_subplot(2, 2, (1,2))
    ax2 = ax.twinx()
    df['A'].plot(ax=ax, color = 'g')
    df['B'].plot(ax=ax2, color ='g')

    fig.add_subplot(223)
    df['C'].plot(color='r')
    plt.axvline(x = 7, color = 'b', label = 'axvline - full height')
    
    fig.add_subplot(2,2,4)
    df = df.round(0)
    ax3 = df['D'].plot(kind='bar')
    ax3.legend( prop={'size': 6})

    fig.tight_layout()
    plt.show()

Which produces this plot: (1,2) in ax = fig.add_subplot(2, 2, (1,2)) does not say location or plot number, isn't it confusing? This is because ax = fig.add_subplot(2, 2, (3,2)) would mean there are 3 plots and taking 2 places but it throws this error

IndexError: GridSpec slice would result in no space allocated for subplot

Why the error?

CodePudding user response：

Your interpretation of the tuple (third parameter) is incorrect: ax = fig.add_subplot(2, 2, (3,2)) says 4 subplots (2 rows x 2 columns) with the following indices:

 --- --- 
| 1 | 2 |
 --- --- 
| 3 | 4 |
 --- ---

and (3, 2) would be a subplot that spans from index # 3 to # 2 which is impossible as they are on a diagonal. (1, 3) would be a subplot on the left over the full hight, (3, 4) at the bottom over the full width etc.

From the docs of subplot:

Three integers (nrows, ncols, index). The subplot will take the index position on a grid with nrows rows and ncols columns. index starts at 1 in the upper left corner and increases to the right. index can also be a two-tuple specifying the (first, last) indices (1-based, and including last) of the subplot, e.g., fig.add_subplot(3, 1, (1, 2)) makes a subplot that spans the upper 2/3 of the figure.