如何用Pandas按FirstName分组并合并唯一Language列表？

Hey there! Let's get your Pandas grouping working correctly. The issue with your current code is that when you use list(set(x)) in the apply function, x is a Series where each element is a single-item list (like ['en']). So set(x) is creating a set of those lists instead of extracting the language strings inside them—definitely not what you want!

正确解法1：先展开列表再分组（推荐，更直观） #

This approach mirrors the logic of your SQL query: first "unpack" the Language lists into individual rows, then group by FirstName, collect distinct languages, and sort them (just like ORDER BY in SQL):

import pandas as pd

# 你的原始数据
items = [
    {'FirstName': 'David', 'Language': ['en',]},
    {'FirstName': 'David', 'Language': ['fr',]},
    {'FirstName': 'David', 'Language': ['en',]},
    {'FirstName': 'Bob', 'Language': ['en',]}
]

df = pd.DataFrame(items)

# 1. 把Language列的列表拆分成单独的行
exploded_df = df.explode('Language')

# 2. 分组后获取去重的语言，排序后转为列表
result_df = exploded_df.groupby('FirstName')['Language'].agg(lambda x: sorted(x.unique())).reset_index()

# 3. 转成你需要的字典数组格式
final_result = result_df.to_dict('records')
print(final_result)

输出结果：

[{'FirstName': 'Bob', 'Language': ['en']}, {'FirstName': 'David', 'Language': ['en', 'fr']}]

正确解法2：直接在分组时拼接列表并去重 #

If you prefer to avoid the explode step, you can use a list comprehension to flatten all the Language lists in each group, then deduplicate and sort:

result_df = df.groupby('FirstName')['Language'].agg(
    lambda x: sorted(list(set(lang for sublist in x for lang in sublist)))
).reset_index()

final_result = result_df.to_dict('records')

为什么你的原始代码不对？ #

Let's break down the problem with your original code:

df.groupby('FirstName')['Language'].apply(lambda x: list(set(x)))

When you group by FirstName, x for the "David" group is a Series like this:

0    [en]
1    [fr]
2    [en]
Name: Language, dtype: object

Using set(x) here creates a set of list objects ({['en'], ['fr']}), not a set of language strings. Converting that to a list gives you something like [['fr'], ['en']]—which is not the flat, deduplicated list you need.

内容的提问来源于stack exchange，提问作者David542