如何基于DataFrame现有列批量生成计算后的新列？

Hey there! Let's get this sorted for you. Your goal is to generate weighted columns for every date column in your sega_df using the formula (100 - [5*eachcell])*0.2, with new columns named Weighted_<original-col-name>, and optionally cap negative values at 0. Here's how to do it efficiently:

First, let's fix your original approach issues #

Your current loop has a couple of problems:

• You're iterating over row indices instead of columns, which isn't what you need
• You keep overwriting a single Weighted column instead of creating new ones for each original column
• The line sega_df[sega_df < 0] = 0 modifies all columns (including your original data) instead of just the new weighted columns

Solution 1: Simple column-wise loop (easy to read) #

This approach loops through each column, calculates the weighted values, and adds the new column directly to your DataFrame:

# Loop through each column in your DataFrame
for col in sega_df.columns:
    # Create the new column name
    weighted_col_name = f"Weighted_{col}"
    # Apply your formula to the column
    sega_df[weighted_col_name] = (100 - 5 * sega_df[col]) * 0.2
    # Optional: Cap negative values at 0 (remove this line if you want to keep negatives)
    sega_df[weighted_col_name] = sega_df[weighted_col_name].clip(lower=0)

Solution 2: Vectorized operation (faster for large datasets) #

If you're working with a big DataFrame, this vectorized method is more efficient—it calculates all weighted columns at once without explicit loops:

import pandas as pd

# Calculate weighted values for all columns in one go
weighted_data = (100 - 5 * sega_df) * 0.2
# Rename columns to add the Weighted_ prefix
weighted_data.columns = [f"Weighted_{col}" for col in weighted_data.columns]
# Optional: Cap negatives at 0 (remove if not needed)
weighted_data = weighted_data.clip(lower=0)
# Merge the original DataFrame with the new weighted columns
sega_df = pd.concat([sega_df, weighted_data], axis=1)

Example Output #

Using your sample data:

# First, recreate your original DataFrame
data = {
    "2016-11-01": [12.0, 5.0],
    "2016-12-01": [3.0, 23.0]
}
sega_df = pd.DataFrame(data, index=["Sonic", "Shadow"])
sega_df.index.name = "Character"

After running either solution (without the clip step to keep negatives), you'll get exactly your desired output:

2016-11-01  2016-12-01  Weighted_2016-11-01  Weighted_2016-12-01
Character                                                                     
Sonic             12.0         3.0                  8.0                 17.0
Shadow             5.0        23.0                 15.0                 -3.0

内容的提问来源于stack exchange，提问作者florence-y