Pandas如何仅替换DataFrame中值为空列表的单元格

原始示例数据 #

测试DataFrame构造代码如下：

import pandas as pd

df = pd.DataFrame({
    'variations_list1': [["yellow","ornage"], ["xl","xxl"], ["Burger","pizza"]],
    'variations_list2': [[], [], ["$25","$30"]]
})

原始数据打印效果：

variations_list1          variations_list2
0  ["yellow","ornage"]                        []
1         ["xl","xxl"]                        []
2  ["Burger","pizza"]               ["$25","$30"]

处理逻辑 #

仅替换variations_list2列中值为空列表的单元格，取当前行variations_list1的列表值填充，非空单元格保留原值。

可直接运行的实现代码 #

通过apply按行遍历判断即可，通过判断列表长度是否为0识别空列表，兼容性最强，不会触发pandas的真值判断报错：

df['variations_list2'] = df.apply(
    lambda row: row['variations_list1'] if len(row['variations_list2']) == 0 else row['variations_list2'],
    axis=1
)

避坑提示 #

• 不要尝试直接用df['variations_list2'] == []做整列条件筛选，列表是可变对象，pandas对存储列表的列做整列等值判断时会抛出ValueError: The truth value of a Series is ambiguous报错。

处理后结果 #

执行代码后打印DataFrame，输出完全符合预期：

variations_list1          variations_list2
0  ["yellow","ornage"]        ["yellow","ornage"]
1         ["xl","xxl"]               ["xl","xxl"]
2  ["Burger","pizza"]              ["$25","$30"]

内容的提问来源于stack exchange，提问作者boyenec