如何获取列中负数最多的组并添加另一个条件来过滤组？

这是对此已接受的 答案 的扩展。

我的数据帧：

import pandas as pd
df = pd.DataFrame(
    {
        'a': [-3, -1, -2, -5, 10, -3, -13, -3, -2, 1, 2, -100],
        'b': [1, 2, 3, 4, 5, 10, 80, 90, 100, 99, 1, 12]
    }
)

预期输出：

      a    b
5    -3   10
6   -13   80
7    -3   90
8    -2  100

逻辑：

a) 选择 a 中最长的负数条纹。|| |b) 例如，如果有两条大小相同的条纹，我想要

总和更大的一条。在 b 中有两条大小为4的条纹，但我想要第二条，因为 df 的总和更大。 b 我的尝试：

My Attempt:

import numpy as np
s = np.sign(df['a'])
df['g'] = s.ne(s.shift()).cumsum()
df['size'] = df.groupby('g')['g'].transform('size')
df['b_sum'] = df.groupby('g')['b'].transform('sum')

import numpy as np

def find_group_with_most_negatives(df):
    """
    Finds the group in DataFrame 'df' with the most negative numbers in column 'a',
    breaking ties by selecting the group with the largest sum in column 'b'.

    Args:
        df (pd.DataFrame): The input DataFrame with columns 'a' and 'b'.

    Returns:
        pd.DataFrame: A subset of the input DataFrame containing the group with the most negatives.
    """
    # Calculate sign changes and group consecutive negative/positive sequences
    s = np.sign(df['a'])
    df['g'] = s.ne(s.shift()).cumsum()

    # Calculate group size and sum of 'b' for each group
    df['size'] = df.groupby('g')['g'].transform('size')
    df['b_sum'] = df.groupby('g')['b'].transform('sum')

    # Find the group with the maximum size (number of negatives)
    max_size_group = df['size'].max()
    df_filtered_size = df[df['size'] == max_size_group]

    # If there are multiple groups with the same maximum size,
    # filter for the group with the maximum sum of 'b'
    if len(df_filtered_size) > 1:
        max_b_sum = df_filtered_size['b_sum'].max()
        df_filtered_size = df_filtered_size[df_filtered_size['b_sum'] == max_b_sum]

    return df_filtered_size[['a', 'b']]

# Apply the function to your DataFrame
result_df = find_group_with_most_negatives(df.copy())

# Print the result
print(result_df)

Explanation:

1. Sign Changes and Grouping: This part remains the same as your attempt. It calculates sign changes in column 'a' and groups consecutive negatives.
2. Group Calculations: Like your attempt, it calculates the size of each group (number of negatives) and the sum of values in column 'b' for each group.
3. Filtering by Maximum Size: It identifies the group with the maximum size.
4. Filtering by Maximum Sum (Tiebreaker): If there are multiple groups with the same maximum size, it further filters these groups to select the one with the maximum sum of 'b' values.
5. Returning the Subset: Finally, it returns the subset of the DataFrame corresponding to the selected group.

This code will efficiently identify the desired group based on the specified criteria.