Key Error:3 while using For to plot using matplotlib.pyplot.scatter

问题

I have a dataset with 10 columns, in which column 4 and 8 are the data I want to plot, and column 9 is determining the marker type.

mkr_dict = {'Fluoride': 'D', 'Chloride': 'P', 'Bromide': 'o', 'Iodide' : '^'}
for kind in mkr_dict:
    d = df[df.Halogen==kind]
    plt.scatter(d[3], d[7], 
                s = 20,
                c = 'red',
                marker=mkr_dict[kind])
plt.show()

I keep getting Key Error: 3 while executing. Can someone give me an advice?

回答1:

As you can see in the following minimal example, the code from the question may run fine.

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

a = np.random.rand(50,8)
df = pd.DataFrame(a)
df["Halogen"] = np.random.choice(['Fluoride','Chloride','Bromide','Iodide'], size=50)


mkr_dict = {'Fluoride': 'D', 'Chloride': 'P', 'Bromide': 'o', 'Iodide' : '^'}
for kind in mkr_dict:
    d = df[df.Halogen==kind]
    plt.scatter(d[3], d[7], 
                s = 20,
                c = 'red',
                marker=mkr_dict[kind])
plt.show()

However, if the columns are not numbered, but have a name instead, something like d[3] will fail. You would then need to select the column by its index,

d.iloc[:,3]

or by its name (in the case below, e.g. d.D or d["D"]). A version using iloc:

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

a = np.random.rand(50,8)
df = pd.DataFrame(a, columns=list("ABCDEFGH"))
df["Halogen"] = np.random.choice(['Fluoride','Chloride','Bromide','Iodide'], size=50)


mkr_dict = {'Fluoride': 'D', 'Chloride': 'P', 'Bromide': 'o', 'Iodide' : '^'}
for kind in mkr_dict:
    d = df[df.Halogen==kind]
    plt.scatter(d.iloc[:,3], d.iloc[:,7], 
                s = 20,
                c = 'red',
                marker=mkr_dict[kind])
plt.show()

来源：https://stackoverflow.com/questions/45263070/key-error3-while-using-for-to-plot-using-matplotlib-pyplot-scatter

标签

python

matplotlib

keyerror