如何处理Pandas DataFrame中的NA值？附待优化代码

原代码存在两个问题：一是无法处理输入中的NA值，二是逻辑错误——仅依据Delivery Time zone overlap列的值就给所有三组输出列赋值，不符合"每个输入列对应三个输出列"的需求。以下是修正后的代码，同时满足NA值处理要求：

测试数据（含NA值） #

import pandas as pd
import numpy as np

# 构造带NA值的输入DataFrame
delivery_df = pd.DataFrame({
    'Delivery Time zone overlap': [3, np.nan, 11.5, 3, 0, 9.5],
    'Developer & Product Time zone overlap': [np.nan, 3, 11.5, 3, 0, 9.5],
    'Developer & Agile champion Time zone overlap': [3, 3, np.nan, 3, 0, 9.5]
})

修正后的代码 #

def process_overlap_column(col_name, df):
    """处理单个输入列，生成对应的三个输出列"""
    col_values = df[col_name]
    base_name = col_name.replace(' Time zone overlap', '')
    
    # 初始化三个输出列，默认设为NA
    colocation_col = f'{base_name} colocation'
    overlap_col = f'{base_name} overlap'
    opposite_col = f'{base_name} opposite timezone'
    
    # 创建结果字典
    result = pd.DataFrame({
        colocation_col: np.nan,
        overlap_col: np.nan,
        opposite_col: np.nan
    }, index=df.index)
    
    # 处理非NA值的情况
    mask_not_na = col_values.notna()
    
    # 条件判断：colocation（值为0）
    mask_coloc = (col_values == 0) & mask_not_na
    result.loc[mask_coloc, colocation_col] = f'{base_name} is colocated'
    
    # 条件判断：overlap（1-5.5之间）
    mask_overlap = (col_values.between(1, 5.5)) & mask_not_na
    result.loc[mask_overlap, overlap_col] = f'{base_name} has overlap'
    
    # 条件判断：opposite timezone（大于5.5）
    mask_opposite = (col_values > 5.5) & mask_not_na
    result.loc[mask_opposite, opposite_col] = f'{base_name} in opposite timezone'
    
    return result

# 处理所有输入列，合并结果
output_dfs = []
for col in delivery_df.columns:
    output_dfs.append(process_overlap_column(col, delivery_df))

output_df = pd.concat(output_dfs, axis=1)

# 打印结果
print(output_df)

关键修改说明 #

• 按列独立处理：每个输入列单独生成对应的三个输出列，不再依赖单一列的值批量赋值
• NA值处理：所有输出列初始化为NA，仅当输入列非NA时，根据值的范围填充对应输出列；若输入列是NA，对应的三个输出列保持NA
• 高效操作：使用Pandas的向量化方法（loc、notna、between）替代iterrows和append，提升运行效率
• 语法修正：原代码中"Developer & Product has overlap"的表述不符合语法，已统一调整为正确的表述逻辑

内容的提问来源于stack exchange，提问作者Leo82