转换数据帧中的测量单位

Question

我有一个数据帧：

df = pd.DataFrame({"id": [1, 2, 3, 4], "unit_measure": ["pound", "gram", "pound", "gram"], "waste": [10,1000,20,1000]})

id   unit_measure    waste  
1       pound          10       
2       gram         1000
3       pound          20
4       gram         1000

我想通过将gram转换为lb来创建一个新列

id   unit_measure   waste   lb_waste
1       pound          10       10
2       gram         1000      2.2  
3       pound          20       20
4       gram         1000      2.2

我应用了以下代码：

def lb(df):
  if df['unit_measure'] == 'pound':
    return df.waste
  else:
    return df['waste']*0.0022
df['lb_waste'] = df.apply(lb, axis=1)

这给了我以下错误：

The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

Answer 1

要修复实现，请引用函数参数，而不是全局df1。在这里，我将参数名称从df更改为row，以更清楚地了解它所代表的内容（@Psidom，我似乎同意这一点）：

def lb(row):
    if row['unit_measure'] == 'pound':
        return row['waste']
    else:
        return row['waste'] * 0.0022


df1['lb_waste'] = df1.apply(lb, axis=1)

---

然而，有许多“熊猫”方法比极慢的apply方法更有效。例如，@Nk03的^{}解决方案

或通过^{}的另一个选项：

df1['lb_waste'] = df1['waste'].where(df1['unit_measure'].eq('pound'),
                                     df1['waste'] * 0.0022)

   id unit_measure  waste  lb_waste
0   1        pound     10      10.0
1   2         gram   1000       2.2
2   3        pound     20      20.0
3   4         gram   1000       2.2

Answer 2

尝试：

import numpy as np
df['lb_waste'] = np.where(df['unit_measure'] == 'pound', df.waste , df['waste']*0.0022)