Pandas技术实现：列值符合条件时取两列最大值否则取均值

Got it, let's figure out how to add that conditional logic to your adj_mean column. You're halfway there with the max function—we just need to layer in the condition for when both values are positive. Here are two solid approaches:

Method 1: Using numpy.where (Best for Large Datasets) #

This uses vectorized operations, which are way faster than row-by-row processing for big data. First, make sure you have numpy imported:

import numpy as np

Then write the conditional logic:

# Check if either Avg or rolling_mean is zero
has_zero = (df['Avg'] == 0) | (df['rolling_mean'] == 0)

# Assign adj_mean: max if there's a zero, average otherwise
df['adj_mean'] = np.where(
    has_zero,
    df[['Avg', 'rolling_mean']].max(axis=1),
    df[['Avg', 'rolling_mean']].mean(axis=1)
)

Method 2: Using pandas.apply (Simpler for Small Data) #

If your dataset isn't huge, a row-wise lambda function is easy to read and implement:

df['adj_mean'] = df.apply(
    lambda row: max(row['Avg'], row['rolling_mean']) 
    if row['Avg'] == 0 or row['rolling_mean'] == 0 
    else (row['Avg'] + row['rolling_mean']) / 2,
    axis=1
)

Let's Test It With Your Sample Data #

Both methods will give you exactly the output you want:

Quick Tips #

• Go with Method 1 if you're working with large datasets—vectorized operations avoid slow loops through each row.
• If your data has floating-point numbers that might be almost zero (due to precision), replace ==0 with something like abs(col) < 1e-9 to catch those edge cases.

内容的提问来源于stack exchange，提问作者TroyMcClure8998