misplaced scatter/hexbin subplot colorbars

[javadnoorb]

• closes xlabel disappears in scatter plots with subplots and colorbar #20455
• tests added / passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

NOTE: I have tested the following in Jupyter lab.

When I run:

%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
random_array = np.random.random((1000,3))
df = pd.DataFrame(random_array,columns=['A label','B label','C label'])

fig, axes = plt.subplots(1, 2)
df.plot.scatter('A label','B label', c='C label', ax=axes[0])
df.plot.scatter('A label','B label', c='C label', ax=axes[1])
plt.tight_layout()

I get:

This PR fixes the issue with colorbars. After the PR the output is:

This also fixes similar issue with hexbin plot. To explore different configurations try the following function:

%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

def make_multiplots(numrows=2,numcols=2,sharey=True,sharex=True, is_scatter = True):
    fig, axes = plt.subplots(numrows, numcols, sharey=sharey, sharex=sharex )
    for i in range(axes.shape[0]):
        for j in range(axes.shape[1]):
            random_array = np.random.random((1000,3))
            df = pd.DataFrame(random_array,columns=['A label','B label','C label'])
            if is_scatter:
                df.plot.scatter('A label', 'B label', c='C label', ax=axes[i,j]);
            else:
                df.plot.hexbin('A label', 'B label', gridsize=20, ax=axes[i,j]);

make_multiplots(numrows=2,numcols=3,sharey=False, sharex=False, is_scatter=False)
plt.tight_layout()

[jreback]

can u modify the test to catch this? (if possible / if not ok)

[jreback]

you don’t need another entry, just add on to the previous one (just the number is ok)

[codecov]

Codecov Report

Merging #21728 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #21728      +/-   ##
==========================================
+ Coverage   91.92%   91.93%   +<.01%     
==========================================
  Files         158      159       +1     
  Lines       49705    49719      +14     
==========================================
+ Hits        45693    45707      +14     
  Misses       4012     4012

Flag	Coverage Δ
#multiple	90.3% <100%> (ø)	⬆️
#single	42.02% <0%> (+0.02%)	⬆️

Impacted Files	Coverage Δ
pandas/plotting/_core.py	83.57% <100%> (ø)	⬆️
pandas/core/arrays/period.py	86.53% <0%> (-13.47%)	⬇️
pandas/core/arrays/datetimelike.py	94.38% <0%> (-0.14%)	⬇️
pandas/core/indexes/timedeltas.py	91.05% <0%> (-0.04%)	⬇️
pandas/core/indexes/datetimes.py	95.48% <0%> (-0.01%)	⬇️
pandas/core/generic.py	96.21% <0%> (-0.01%)	⬇️
pandas/core/arrays/timedelta.py	100% <0%> (ø)	⬆️
pandas/core/arrays/datetimes.py	100% <0%> (ø)	⬆️
pandas/_libs/tslibs/__init__.py	100% <0%> (ø)
pandas/core/indexes/datetimelike.py	96.98% <0%> (+0.11%)	⬆️
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 62a0ebc...ad2b52a. Read the comment docs.

[javadnoorb]

Thanks @jreback. I included a test that checks if the distance between subplot parent axes is almost the same as the distance between colorbars.

I'm not sure what @pytest.mark.slow does. I assumed it's for skipping tests that are slow. Do I need it?

[jreback]

thanks @javadnoorb