如何在Pandas中替换DataFrame连续四列的负值为指定值？

Hey there! Let's fix this problem properly—Pandas is designed to avoid those messy nested loops, so we can do this way more cleanly.

最简洁的Pandas原生方法 #

First, target the specific columns you care about (T1 to T4) and use either mask() or where() to replace negative values in one line:

Using mask() #

This replaces values where the condition is True (i.e., values < 0) with your desired value (-5):

# Define the columns we want to modify
target_cols = ['T1', 'T2', 'T3', 'T4']
# Replace negatives with -5
df[target_cols] = df[target_cols].mask(df[target_cols] < 0, -5)

Using where() #

This does the opposite: keeps values where the condition is True (values >= 0), and replaces the rest with -5:

df[target_cols] = df[target_cols].where(df[target_cols] >= 0, -5)

Alternative: Using applymap() #

If you prefer a more explicit function-based approach, applymap() works too (great for custom logic if you ever need it):

df[target_cols] = df[target_cols].applymap(lambda x: -5 if x < 0 else x)

Why your original code didn't work #

Let's break down the issues with your loop approach:

• Your outer loop for i in df.iloc[:, ...]<0 is iterating over column labels of the boolean mask, not the actual data rows/values.
• When you do for j in df[i], j is just a copy of the value from the DataFrame—changing j = -5 doesn't modify the original DataFrame at all (you're not updating the actual cells).
• Nested loops are inefficient in Pandas, especially with larger datasets—vectorized operations (like the methods above) are way faster and more readable.

Result after running the code #

Your DataFrame will look like this:

内容的提问来源于stack exchange，提问作者hegdep