如何在Pandas DataFrame列名上方添加多行表头信息？

Absolutely—you can do this cleanly using Pandas' built-in methods, no post-processing of CSV files required. The key idea is to create a dummy DataFrame with your 4 rows of NaNs, export it first (without headers), then append your original DataFrame (including its header) to the same output file.

Step-by-Step Implementation #

First, import the necessary modules and define your sample DataFrame:

import pandas as pd
import numpy as np

# Your original DataFrame
data = {'A': ['val', 'val'], 'B': ['val', 'val'], 'C': ['val', 'val']}
df = pd.DataFrame(data)

Next, create a dummy DataFrame with 4 rows of NaNs that matches the column count of your original DataFrame:

# Create 4 rows of NaNs with identical columns to your original DataFrame
num_blank_rows = 4
dummy_df = pd.DataFrame(np.full((num_blank_rows, len(df.columns)), np.nan), columns=df.columns)

Finally, export the dummy rows first (without headers), then append your original DataFrame (with its header) to the same file:

# Export dummy rows (no header, space-separated as in your example)
dummy_df.to_csv('output.txt', index=False, header=False, sep=' ')

# Append original DataFrame with its header
df.to_csv('output.txt', index=False, header=True, sep=' ', mode='a')

What This Produces #

The output file will exactly match your desired format:

nan nan nan
nan nan nan
nan nan nan
nan nan nan
A B C
val val val
val val val

Notes for Flexibility #

• Separator Adjustments: If you need a different separator (like commas for standard CSV), just modify the sep parameter (e.g., sep=',').
• Large Files: This approach is efficient even for large datasets because Pandas writes incrementally—you won’t need to load the entire file into memory for post-processing.
• In-Memory Buffer: If you want to work with the output in memory first (instead of writing directly to a file), use StringIO:

  from io import StringIO
  
  buffer = StringIO()
  dummy_df.to_csv(buffer, index=False, header=False, sep=' ')
  df.to_csv(buffer, index=False, header=True, sep=' ')
  
  # Retrieve the full content
  output_content = buffer.getvalue()
  

内容的提问来源于stack exchange，提问作者komodovaran_