问题
Have the following piece of code through which I am trying to plot a graph:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import mpld3
my_list = [1,2,3,4,5,7,8,9,11,23,56,78,3,3,5,7,9,12]
new_list = pd.Series(my_list)
df1 = pd.DataFrame({'Range1':new_list.value_counts().index, 'Range2':new_list.value_counts().values})
df1.sort_values(by=["Range1"],inplace=True)
df2 = df1.groupby(pd.cut(df1["Range1"], [0,1,2,3,4,5,6,7,8,9,10,11,df1['Range1'].max()])).sum()
objects = df2['Range2'].index
y_pos = np.arange(len(df2['Range2'].index))
plt.bar(df2['Range2'].index.values, df2['Range2'].values)
but getting the following error message:
TypeError: float() argument must be a string or a number, not 'pandas._libs.interval.Interval'
Not getting from where this float error is coming. Any suggestion is highly appreciated.
回答1:
Matplotlib cannot plot category datatypes. You would need to convert to a string.
plt.bar(df2['Range2'].index.astype(str), df2['Range2'].values)
回答2:
The pd.cut operation yields intervals:
In [11]: pd.cut(df1["Range1"], [0,1,2,3,4,5,6,7,8,9,10,11,df1['Range1'].max()])
Out[11]:
12 (0, 1]
11 (1, 2]
0 (2, 3]
10 (3, 4]
3 (4, 5]
2 (6, 7]
9 (7, 8]
1 (8, 9]
8 (10, 11]
7 (11, 78]
5 (11, 78]
4 (11, 78]
6 (11, 78]
Name: Range1, dtype: category
Categories (12, interval[int64]): [(0, 1] < (1, 2] < (2, 3] < (3, 4] ... (8, 9] < (9, 10] < (10, 11] <
(11, 78]]
When used in the groupby operation, they are matched based on the index of the cut operation above, and then grouped and summed according to the operation you specified.
As a result, the intervals end up as the index in df2:
In [14]: df2
Out[14]:
Range1 Range2
Range1
(0, 1] 1 1
(1, 2] 2 1
(2, 3] 3 3
(3, 4] 4 1
(4, 5] 5 2
(5, 6] 0 0
(6, 7] 7 2
(7, 8] 8 1
(8, 9] 9 2
(9, 10] 0 0
(10, 11] 11 1
(11, 78] 169 4
When you use df2['Range2'].index.values it will be an array of these intervals passed as the first argument to bar, which is not convertible to a float in the way matplotlib expects.
If you are looking to just plot a bar chart of df2.Range2 and you are happy to have the intervals as the axis labels, this will work:
plt.bar(range(len(df2)), df2.Range2.values, tick_label=df2.Range2.index.values)
and produces this image for me:
来源：https://stackoverflow.com/questions/53619746/python-float-argument-must-be-a-string-or-a-number-not-pandas