GroupBy 和聚合集合 [英] GroupBy and aggregate with set intersection

问题描述

我有一个带有集合列的 Pandas DataFrame:

I have a pandas DataFrame with a sets column:

import pandas as pd

df = pd.DataFrame({'group_var': [1,1,2,2], 'sets_var': [set([0, 1]), set([1, 2]), set([3, 4]), set([5, 6, 7])]})
df

   group_var sets_var
0          1      {0, 1}
1          1      {1, 2}
2          2      {3, 4}
3          2   {5, 6, 7}

我希望对 group_var 进行 groupby 并得到所有对应的 sets_var 集合的交集，如下所示:

I wish to groupby the group_var and get the intersection of all corresponding sets of sets_var, like so:

   group_var sets_var
0          1      {1}
1          2      {}

或像这样的系列:

   sets_var
1  {1}
2  {}

我将如何优雅地完成它?性能是重中之重.

How would I go about it in elegance? Performance is top priority.

推荐答案

使用groupby、agg，并使用set.intersection减少.

df.groupby('group_var', as_index=False).agg(lambda x: set.intersection(*x))

   group_var sets_var
0          1      {1}
1          2       {}

如果性能绝对重要，我们可以尝试去掉lambda:

If performance is absolutely important, we can try getting rid of the lambda:

from functools import partial, reduce 
import operator

p = partial(reduce, operator.and_)
df.groupby('group_var', as_index=False).agg(p)

   group_var sets_var
0          1      {1}
1          2       {}

但是，这仅执行成对交叉，因此您的里程可能会有所不同.

However, this only performs a pairwise intersection, so your mileage may vary.

或者，作为一个系列，

pd.Series({
    k: set.intersection(*g.tolist()) 
    for k, g in df.groupby('group_var')['sets_var']})

1    {1}
2     {}
dtype: object

这篇关于GroupBy 和聚合集合的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！