如何在Pandas中仅查看列子集并保留省略号以适配论文展示？

Got it, I totally get where you're coming from—when you've got an 80-column DataFrame, even Pandas' default truncated view can still stretch too wide to easily pop into a paper. The good news is you can tweak the display settings to show fewer columns while keeping that handy [...] ellipsis, so it fits your page perfectly. Here's how:

1. Tweak Global Pandas Display Settings #

This changes the default view for all DataFrames in your notebook, which is great if you want consistent compact output. Just set display.max_columns to the total number of columns you want visible (Pandas will split this evenly between the start and end of your columns, with an ellipsis in the middle):

import pandas as pd

# Create a test 80-column DataFrame (replace with your actual data)
df = pd.DataFrame({f'Column_{i+1}': range(10) for i in range(80)})

# Set max visible columns to 8 (shows first 4 + ... + last 4)
pd.set_option('display.max_columns', 8)

# Now view your DataFrame—compact and ready to copy!
display(df)

You can also pair this with display.width to cap the total character width of the output, which helps if your column names are long:

pd.set_option('display.width', 100)  # Limit total output width to 100 characters
pd.set_option('display.max_columns', 6)

2. Temporary Display Adjustments (No Global Changes) #

If you don't want to mess with global settings for all your DataFrames, use pd.option_context() to apply the tweak only for a specific block of code. This is perfect for when you just need one compact view for your paper:

# Use a context manager to temporarily set max columns
with pd.option_context('display.max_columns', 6):
    display(df)  # Shows first 3 + ... + last 3 columns

# Or generate a compact string directly, ready to copy-paste
compact_output = df.to_string(max_cols=6)
print(compact_output)

3. Custom Column Selection (If You Want Exact Control) #

If you want to pick exactly which columns show up (e.g., first 2, last 2, with an ellipsis), you can manually construct a view. Note: This won't show the automatic ellipsis, but you can add a dummy column to mimic it if needed:

# Pick first 2 and last 2 columns
trimmed_df = df.iloc[:, :2].join(df.iloc[:, -2:])
# Add a dummy "..." column to match the default look
trimmed_df.insert(2, "...", ["..."]*len(trimmed_df))

display(trimmed_df)

This last method is a bit more manual, but it lets you tailor the exact columns and the ellipsis placement to your paper's needs.

Whichever method you choose, you'll get that clean, truncated view you love—just narrow enough to fit your page and easy to copy into your document.

内容的提问来源于stack exchange，提问作者t.pellegrom