Python拼接两个函数的字符串输出遇TypeError问题求助

错误原因 #

你的代码报错是因为DataFrame中存在pd.NA缺失值，当你在apply的函数里执行df['NLT'] != ""时，如果df['NLT']是NA，这个判断会返回pd.NA，而Python的if语句无法识别NA的布尔值，因此抛出boolean value of NA is ambiguous错误。即使你将字段转为string类型，缺失值依然会以pd.NA的形式存在，和空字符串""是不同的概念。

解决方案 #

方案1：修复apply函数，显式处理NA值 #

在判断空字符串前，先检查是否为NA，用pd.isna()检测缺失值：

import pandas as pd

def a(df):
    if not pd.isna(df['NLT']) and df['NLT'] != "":
        return df['NLT']
    else:
        return df['LT']

def b(df):
    if not pd.isna(df['NCC']) and df['NCC'] != "":
        return df['NCC']
    else:
        return df['CC']

df['ra'] = df.apply(a, axis=1)
df['rb'] = df.apply(b, axis=1)
df['RR'] = df['ra'] + df['rb']

方案2：用矢量化操作替代apply（推荐，更高效） #

pandas的矢量化函数比apply性能更好，适合批量处理DataFrame：

方法A：使用DataFrame.where

# 优先取非空非NA的NLT，否则取LT
df['ra'] = df['NLT'].where((~df['NLT'].isna()) & (df['NLT'] != ""), df['LT'])
# 优先取非空非NA的NCC，否则取CC
df['rb'] = df['NCC'].where((~df['NCC'].isna()) & (df['NCC'] != ""), df['CC'])
# 拼接结果
df['RR'] = df['ra'] + df['rb']

方法B：使用numpy.where

import numpy as np

df['ra'] = np.where((~df['NLT'].isna()) & (df['NLT'] != ""), df['NLT'], df['LT'])
df['rb'] = np.where((~df['NCC'].isna()) & (df['NCC'] != ""), df['NCC'], df['CC'])
df['RR'] = df['ra'] + df['rb']

方案3：将NA替换为空字符串 #

如果你的数据中所有缺失值都应该是空字符串，可以先统一替换：

# 将指定列的NA替换为空字符串
df[['NLT', 'LT', 'NCC', 'CC']] = df[['NLT', 'LT', 'NCC', 'CC']].fillna("")
# 之后原代码即可正常运行
df['ra'] = df.apply(a, axis=1)
df['rb'] = df.apply(b, axis=1)
df['RR'] = df['ra'] + df['rb']

验证示例数据结果 #

用你提供的数据集处理后，RR列的结果为：

内容的提问来源于stack exchange，提问作者Horse_Face_Billy12