在python pandas中根据if-elif-else条件创建一个新列

Question

I want to add a d column to a DataFrame. When creating the d column, the conditions are as follows:

If it starts with 11 in column a , it gets the value in column b , otherwise it gets the value in column c .

a     b    c
11101 100  150
11201 200  250
13301 300  350


a     b    c    d
11101 100  150  100
11201 200  250  200
13301 300  350  350

Answer 1

Use np.where after converting the column a into str :

df["a"] = df["a"].astype("str")
df["d"] = np.where(df["a"].str.startswith("11"), df["b"], df["c"])

or if you don't want to change datatype of a in original dataframe

df["d"] = np.where(df["a"].astype("str").str.startswith("11"), df["b"], df["c"])

Answer 2

您还可以使用 lambda 函数：

df["d"] = df.apply(lambda x : x["b"] if str(x["a"])[:2] == "11" else x["c"], axis =1)

Answer 3

df["d"]=0 for(i , x in enumerate(df["a"])): if str(x).startswith("11"): df["d"][i]=df["b"][i] else: df["d"][i]=df["c"][i] print(df)

Answer 4

To keep it readable, I would prefer the use of an external function ( calculateValue ) in the lambda function

def calculateValue(a, b, c):
  if(a.startswith('11')):
    return b
  else:
    return c

df["d"] = df.apply(lambda row: calculateValue(row["a"],row["b"],row["c"]), axis=1)

Answer 5

You can also use numpy.vectorize which can be a bit faster.

import numpy as np


def get_d(a, b, c):
    return b if str(a)[:2] == '11' else c


df['d'] = np.vectorize(get_d)(df.a, df.b, df.c)