Python Pandas Plot has Gaps when using Sparse Dataframe-CodePudding

When I try to plot the following data

import pandas as pd
dfx = pd.DataFrame()
dfx['bing'] = [1,2,3,np.nan,np.nan,np.nan,np.nan,np.nan,np.nan]
dfx['bang'] = [np.nan,np.nan,np.nan,4,5,6,np.nan,np.nan,np.nan]
dfx['boo'] = [np.nan,np.nan,np.nan,np.nan,np.nan,np.nan,7,8,9]
dfx.plot()
plt.show()

I get this line with gaps. I want to be able to make use of the different colors for different segments of my line, but would like it to be "connected" as well.

CodePudding user response：

Two possible ideas:

Option 1: Use

Option 2: Add line to connect the other lines

A bit more work (though there may be a quicker way to construct the "connector" line than the approach I'm using here). Add a column that connects end of one line with start of the next. E.g.:

dfx = pd.DataFrame()
dfx['bing'] = [1,2,3,np.nan,np.nan,np.nan,np.nan,np.nan,np.nan]
dfx['bang'] = [np.nan,np.nan,np.nan,4,5,6,np.nan,np.nan,np.nan]
dfx['boo'] = [np.nan,np.nan,np.nan,np.nan,np.nan,np.nan,7,8,9]

# get index vals end-start: array([2, 3, 5, 6], dtype=int64)
s = dfx.apply(lambda col: (col.first_valid_index(), 
                           col.last_valid_index()))\
    .stack().sort_values().to_numpy()[1:-1]

# create extra col with connection btw consecutive lines
dfx['connect'] = dfx.fillna(method='ffill', axis=1).iloc[:,-1]\
    .mask(~dfx.index.isin(s))

# print(dfx.connect.to_numpy())
# [nan nan  3.  4. nan  6.  7. nan nan]

fig, ax = plt.subplots(figsize=(12,8))

colors = ['blue','orange','green','black']
for i, col in enumerate(dfx.columns):
    if col == 'connect':
        # add `label=col` if you want "connect" in the legend as well
        ax.plot(dfx[col], colors[i], linestyle='dotted')
    else:
        ax.plot(dfx[col], colors[i], label=col)
ax.legend()
plt.show()

Result: