python 使用多个条件创建一个新列

python create a new column using multiple conditions

我刚从 Python 开始,我有一大堆受试者及其 (BMI) 体重指数(以及更多数据)。 我需要创建一个新列(称为 OMS),我可以在其中声明它们是否为 "normal"、"overweight"、"obese"、等等。

但我就是找不到正确的方法。我尝试了 np.when,但这只适用于 2 个条件。

我尝试了 if、elif 和 else 都没有成功,还有:

df['oms'] = np.nan

df['oms'].loc[(df['IMC'] <=18.5 )] = "slim"

df['oms'].loc[(df['IMC'] >= 18.5) & (df['IMC'] <25 )] = "normal"

df['oms'].loc[(df['IMC'] >= 25) & (df['IMC'] <=30 )] = "overweight"

df['oms'].loc[(df['IMC'] > 30)] = "obese"

有什么想法吗?我卡住了。

df.loc[df['IMC'].lt(18.5), 'oms'] = "slim"
df.loc[df['IMC'].ge(18.5) & df['IMC'].lt(25), 'oms'] = "normal"
df.loc[df['IMC'].ge(25) & df['IMC'].lt(30), 'oms'] = "overweight"
df.loc[df['IMC'].ge(30), 'oms'] = "obese"

您也可以使用pd.cut

bins = [0, 18.5, 25, 30, 9999]
labels = ['slim', 'normal', 'overweight', 'obese']

df = pd.DataFrame({'IMC': [15, 20, 27, 40]})
df['oms'] = pd.cut(df['IMC'], bins, labels=labels)
>>> df
   IMC         oms
0   15        slim
1   20      normal
2   27  overweight
3   40       obese

也许试试:

df['oms'] = ""#keep it object dtype

df.loc[(df['IMC'] <=18.5 ), 'oms'] = "slim"
df.loc[(df['IMC'] >= 18.5) & (df['IMC'] <25 ), 'oms'] = "normal"
df.loc[(df['IMC'] >= 25) & (df['IMC'] <=30 ), 'oms'] = "overweight"
df.loc[(df['IMC'] > 30), 'oms'] = "obese"

您可以将 lambda 函数和 apply 与熊猫数据框一起使用。

我创建了一个虚拟数据文件:

bmi,height
20,72
22,73
26,77
5,66
13,60

导入数据文件

df = pd.read_csv('data.txt', header=0)

创建了一个列,就像您对 NaN 所做的那样(但您不必这样做)

df["oms"] = np.nan

然后使用 lambda 将 'bmi' 列与某些条件进行比较

df['oms'] = df['bmi'].apply(lambda x: 'slim' if x < 18.5 else ('normal' if x<25 else ('overweight' if x<30 else 'obese')))

数据是这样的,

print(df.head())

   bmi  height     oms
0   20      72  normal
1   22      73   obese
2   26      77   obese
3    5      66  skinny
4   13      60  skinny

使用 numpy.select,我喜欢这个替代方案,因为它用途广泛,您可以轻松添加或删除条件。

import numpy as np

condlist = [df["IMC"] <= 18,
           (df["IMC"] >= 18.5) & (df['IMC'] <25),
           (df["IMC"] >= 25) & (df['IMC'] <=30),
            df["IMC"] > 30]

condchoice = ["slim", "normal", "overweight", "obese"]

df["oms"] = np.select(condlist, condchoice)