How to combine Line PLots & legends in seaborn/matplotlib-CodePudding

I have a dataframe like this

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

df = pd.DataFrame({'A': [1,2,3,4,5,4,3,2,1],
                   'B': [200,300,400,800,1500,900,450,275,200],
                   'c': ['up','up','down','up','up','down','down','up','down'],
                   'd': ['r','r','r','r','r','r','r','r','r'],
                   })

When I do a line plot with seaborn with the code sns.lineplot(data=df, x='A', y='B',ci=None). I get a plot like below.

As you can see, the points have not completely plotted. When I do the plot with the code below

upDF = df.loc[df['c']=='up']
downDF = df.loc[df['c']=='down']
sns.lineplot(data=upDF, x='A', y='B',ci=None, label = 'up');
sns.lineplot(data=downDF, x='A', y='B',ci=None, label = 'down');
plt.legend()

I get a plot like below

In this plot, all points are plotted, but they are 'disconnected'. How can I get a plot like below with the points connected & just only one legend (column 'D' or a constant can be used for legend)

If I use sns.lineplot on my actual data, I get the graph like below instead of up & down plotted separately.

CodePudding user response：

You can use pure matplotlib. Select the rows after an up as part of the up Series to ensure joining the points. For the legend, add a label that starts with an underscore to be skipped:

import matplotlib.pyplot as plt
m = df['c']=='up'
upDF = df.loc[m|m.shift()]
downDF = df.loc[~m]
ax = plt.subplot()
upDF.plot(x='A', y='B', label='r', ax=ax)
downDF.plot(x='A', y='B', label='_skip', ax=ax)
ax.legend()

output: