输入类型不支持ufunc“除法”……尝试获取 NumPy 平均值时出现错误问题

Question

我是 Numpy 的新手，我一直在尝试获取从另一个数组派生的数组的平均值。

这是给我错误的代码：“输入类型不支持 ufunc 'divide'，并且根据转换规则 ''safe''，无法将输入安全地强制为任何支持的类型”

import numpy as np
import pandas as pd
​
cars = pd.read_csv('data/co2_emissions_canada.csv')
cars_makes = cars['Make'].to_numpy()
cars_models = cars['Model'].to_numpy()
cars_classes = cars['Vehicle Class'].to_numpy()
cars_engine_sizes = cars['Engine Size(L)'].to_numpy()
cars_cylinders = cars['Cylinders'].to_numpy()
cars_transmissions = cars['Transmission'].to_numpy()
cars_fuel_types = cars['Fuel Type'].to_numpy()
cars_fuel_consumption = cars['Fuel Consumption Comb (L/100 km)'].to_numpy()
cars_co2_emissions = cars['CO2 Emissions(g/km)'].to_numpy()
​
#the median of the cars_engine_sizes
print(np.median(cars_engine_sizes))

#the average fuel consumption for regular gasoline (Fuel Type = X), #premium gasoline (Z), ethanol (E), and diesel (D)? 

fuel_typesx=np.array(cars_fuel_types[cars_fuel_types=='X'])
print(np.average(fuel_typesx))

fuel_typesz=np.array(cars_fuel_types[cars_fuel_types=='Z'])
print(np.average(fuel_typesz))

fuel_typese=np.array(cars_fuel_types[cars_fuel_types=='E'])
print(np.average(fuel_typese))

请问我缺少什么

Answer 1

我猜完整的错误消息看起来像这样：

In [753]: np.average(np.array(['A','B','C','A'],dtype=object))
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Input In [753], in <cell line: 1>()
----> 1 np.average(np.array(['A','B','C','A'],dtype=object))

File <__array_function__ internals>:5, in average(*args, **kwargs)

File ~\anaconda3\lib\site-packages\numpy\lib\function_base.py:380, in average(a, axis, weights, returned)
    377 a = np.asanyarray(a)
    379 if weights is None:
--> 380     avg = a.mean(axis)
    381     scl = avg.dtype.type(a.size/avg.size)
    382 else:

File ~\anaconda3\lib\site-packages\numpy\core\_methods.py:191, in _mean(a, axis, dtype, out, keepdims, where)
    189         ret = ret.dtype.type(ret / rcount)
    190 else:
--> 191     ret = ret / rcount
    193 return ret

TypeError: ufunc 'true_divide' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''

cars_fuel_types

来自数据帧，显然包含“E”等字符串。 所以它是

object

dtype。 即使您选择相似的值，也不能取“平均值”。

average

取值的总和并除以计数。

sum

对于 Python 字符串来说是连接，而不是某种数学。

In [754]: np.sum(np.array(['A','B','C','A'],dtype=object))
Out[754]: 'ABCA'

Answer 2

我发现使用

evidently.ai

时的 UFuncTypeError 与作为 reference_data 或 current_data 传递的数据帧的索引奇怪地相关。如果您的数据帧有一个命名索引，它将导致错误：“UFuncTypeError：ufunc'multiply'不包含具有签名匹配类型的循环（dtype（'None”

）

解决方案：要解决此问题，请从数据帧中删除（删除）索引：

x = df.copy()
x.reset_index(drop=True, inplace=True) # <- remove index
report = Report(metrics=[ColumnDriftMetric(column_name="premium")]) # 'premium' is an arbitrary feature in my dataset
report.run(reference_data=x, current_data=x) # <- note: you should set reference_data and current_data accordingly 
report

希望这有帮助！