Python中如何为groupby的agg函数传递随机变量？

架构老林

2026-5-29

How to Use a Random Number in Pandas Groupby Aggregation

Absolutely, you can replace that fixed divisor with a random number—though the string formatting approach you tried won’t work directly, since Pandas’ agg() expects valid aggregation functions or expressions, not interpolated strings. Let’s break down the correct ways to implement this:

Option 1: Use a Single Random Number for All Groups

If you want one consistent random value to apply to every group’s count calculation, first generate the random number outside the agg() call, then reference it in a lambda function:

import pandas as pd
import random

# Sample DataFrame
df = pd.DataFrame({'a': ['x', 'x', 'y', 'y', 'y'], 'b': [1, 2, 3, 4, 5]})

# Generate your random number (adjust the range as needed)
# Use uniform for floats, randint for integers
random_divisor = random.uniform(1, 10)  # Example: random float between 1 and 10

# Perform groupby and aggregation
result = df.groupby('a')['a'].agg(lambda x: x.count() / random_divisor)

This works because the lambda function captures the pre-generated random_divisor and applies it to each group’s count result.

Option 2: Use a Unique Random Number Per Group

If you want a different random divisor for every group, generate the random number directly inside the lambda (this will create a new random value for each group):

result = df.groupby('a')['a'].agg(lambda x: x.count() / random.uniform(1, 10))

Why Your Original Approach Failed

The syntax count('a')/{}.format(random_number) tries to create a string, but Pandas doesn’t interpret interpolated strings as valid aggregation logic. agg() accepts: