如何在Pandas DataFrame中为列分配可迭代对象？修复字典列NaN问题

原始数据与需求 #

你有如下金融股票DataFrame，希望新增cashflows列，每行单元格是基于Characteristic1和Characteristic3生成的现金流字典：

import pandas as pd

df = pd.DataFrame(
    {
        "stockprice": [100, 103, 240],
        "Characteristic1": [1, 3, 3],
        "Characteristic2": [5, 7, 1],
        "Characteristic3": [1, 4, 6],
    },
    index=["Company A", "Company B", "Company C"],
)

错误原因分析 #

你尝试用assign直接传入Series调用cashflow_series函数时，函数接收的是整个Series而非每行的单个值，无法生成对应每行的字典，最终导致列值全为NaN。

修复方案 #

方案1：使用apply逐行处理 #

通过apply指定axis=1，让函数逐行接收每行数据，提取对应特征值生成字典：

def cashflow_series(ch1=1, ch2=2):
    return {0: ch1, 0.5: ch2, 1: 7, 2: 8, 3: 9}

# 新增cashflows列
df['cashflows'] = df.apply(lambda row: cashflow_series(ch1=row['Characteristic1'], ch2=row['Characteristic3']), axis=1)

执行后得到预期结果：

stockprice  Characteristic1  Characteristic2  Characteristic3               cashflows
Company A         100                1                5                1  {0: 1, 0.5: 1, 1: 7, 2: 8, 3: 9}
Company B         103                3                7                4  {0: 3, 0.5: 4, 1: 7, 2: 8, 3: 9}
Company C         240                3                1                6  {0: 3, 0.5: 6, 1: 7, 2: 8, 3: 9}

方案2：修改函数支持批量生成字典 #

如果数据集较大，批量处理效率更高，可修改函数接收Series并通过列表推导式生成每行的字典：

def cashflow_series(ch1_series, ch2_series):
    return [{0: ch1, 0.5: ch2, 1: 7, 2: 8, 3: 9} for ch1, ch2 in zip(ch1_series, ch2_series)]

# 使用assign新增列
df = df.assign(cashflows=cashflow_series(df['Characteristic1'], df['Characteristic3']))

此方法直接基于Series批量生成字典列表，避免逐行循环，性能更优。

内容的提问来源于stack exchange，提问作者Tim