Pandas筛选DataFrame中仅包含3和4的指定分组

Got it, let's work through this problem together. You need to keep only the groups that have both 3 and 4 in col1, with no other values at all—so groups B and D are the ones we want to retain from your sample data.

Step 1: Set Up the Sample Data #

First, let's create the DataFrame with your example values so we can test our solution directly:

import pandas as pd

data = {
    'groups': ['A', 'A', 'A', 'A', 'B', 'B', 'B', 'C', 'D', 'D'],
    'col1': [3, 4, 2, 1, 3, 3, 4, 2, 4, 3]
}
df = pd.DataFrame(data)

Step 2: Filter Valid Groups #

There are a couple of clean ways to do this. Let's start with the most efficient one:

Method 1: Use Groupby + Set Aggregation

This approach leverages set operations to check if a group's values exactly match {3,4}:

# Get the set of unique values for each group
group_values = df.groupby('groups')['col1'].agg(set)

# Filter groups where the set is exactly {3,4}
valid_groups = group_values[group_values == {3, 4}].index

# Keep only rows from valid groups in the original DataFrame
result = df[df['groups'].isin(valid_groups)]

Method 2: Use Groupby Filter with Lambda (More Explicit)

If you prefer a more verbose, readable approach, you can use filter() with a lambda that checks three clear conditions:

• The group contains 3
• The group contains 4
• All values in the group are either 3 or 4

valid_groups = df.groupby('groups').filter(
    lambda x: (3 in x['col1'].values) 
    and (4 in x['col1'].values) 
    and (x['col1'].isin([3, 4]).all())
).groups.keys()

result = df[df['groups'].isin(valid_groups)]

Step 3: Check the Result #

Either method will give you the expected output:

groups  col1
4      B     3
5      B     3
6      B     4
8      D     4
9      D     3

Both methods work, but Method 1 is faster for large datasets since set aggregation is more efficient than row-wise checks in the lambda.

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