Pandas中DataFrame堆叠方法咨询：将含列表值的DataFrame转指定格式

开发者阿泽

2026-5-25

Solution to Stack Pandas DataFrame with List Columns

Hey there! Let's break down how to convert your DataFrame into the desired format. Here's a step-by-step approach using Pandas:

Step 1: Create the Sample DataFrame (for reference)

First, let's replicate your input DataFrame to work with:

import pandas as pd

data = {
    'id': ['id01', 'id02', 'id03', 'id04'],
    'val': [['a', 'b'], ['b'], [], ['a', 'c']]
}
df = pd.DataFrame(data)

Step 2: Filter Out Rows with Empty Lists

We need to drop rows where the val column has an empty list. You can do this by checking the length of each list:

# Remove rows with empty lists in 'val'
df_filtered = df[df['val'].apply(len) > 0]

Alternatively, using str.len() works too (Pandas handles list columns with string methods in newer versions):

df_filtered = df[df['val'].str.len() > 0]

Step 3: Convert List Column to Multiple Columns

Now, turn each element in the val list into a separate column. We'll create a new DataFrame from the list values, using the id column as the index:

# Convert list column to multiple columns
result = pd.DataFrame(df_filtered['val'].tolist(), index=df_filtered['id'])

Step 4: (Optional) Reset Index to Make 'id' a Column

If you prefer id to be a regular column instead of the index, reset it:

result = result.reset_index().rename(columns={'index': 'id'})

Final Output

After running these steps, your result will look like this (with id as index):

0    1
id01  'a'  'b'
id02  'b'  NaN
id04  'a'  'c'

Or if you reset the index:

id    0    1
0  id01  'a'  'b'
1  id02  'b'  NaN
2  id04  'a'  'c'

Quick Extra Tips

To rename numeric columns to something more meaningful (like val1, val2):
```
result.columns = [f'val{i+1}' for i in result.columns]
```
To replace NaN values with empty strings (if you don't want missing values shown):
```
result = result.fillna('')
```