• 假设您在data_info（字典列表）中具有有效的数据结构。*

您可以使用pd.json_normalize从data_info列中提取数据：

# Create an index to allow merging the area column
idx = df.index.repeat(df['data_info'].str.len())

# Extract data and set the area column
dat = (pd.json_normalize(df['data_info'].explode())
         .assign(area=df.loc[idx, 'area'].values))

# Compute the sum
dat = dat.groupby(['color', 'area']).sum().reset_index('color')

# Create the expected output
out = (pd.DataFrame({'area': dat.index, 'data_info': dat.to_dict('records')})
         .groupby('area', as_index=False).agg(list))

字符串
输出量：

>>> out
  area  data_info
0   NY  [{'color': 'Black', 'count': 2, 'weight': 15}, {'color': 'Blue', 'count': 7, 'weight': 17}, {'color': 'Red', 'count': 7, 'weight': 3}, {'color': 'Yellow', 'count': 3, 'weight': 5}]
1   SF  [{'color': 'Green', 'count': 1, 'weight': 14}, {'color': 'Yellow', 'count': 9, 'weight': 2}]

型
中间步骤

>>> (pd.json_normalize(df['data_info'].explode())
       .assign(area=df.loc[idx, 'area'].values))

    color  count  weight area
0  Yellow      3       5   NY
1    Blue      2      11   NY
2     Red      7       3   NY
3   Green      1      14   SF
4  Yellow      9       2   SF
5    Blue      5       6   NY
6   Black      2      15   NY

>>> dat.groupby(['color', 'area']).sum().reset_index('color')
       color  count  weight
area                       
NY     Black      2      15
NY      Blue      7      17
SF     Green      1      14
NY       Red      7       3
NY    Yellow      3       5
SF    Yellow      9       2

型

经典的for循环方式。它在结尾创建一个文件（result.txt）以更好地查看结果。

验证码：

import pandas as pd
data = {
    'ID': ['a001', 'a002', 'a003'],
    'area': ['NY', 'SF', 'NY'],
    'data_info': [
        [{'color': 'Yellow', 'count': 3, 'weight': 5}, {'color': 'Blue', 'count': 2, 'weight': 11}, {'color': 'Red', 'count': 7, 'weight': 3}],
        [{'color': 'Green', 'count': 1, 'weight': 14}, {'color': 'Yellow', 'count': 9, 'weight': 2}],
        [{'color': 'Blue', 'count': 5, 'weight': 6}, {'color': 'Black', 'count': 2, 'weight': 15}]
    ]
}
df = pd.DataFrame(data)

def combine_data_info(group):
    combined_data = {}
    for _, row in group.iterrows():
        for entry in row['data_info']:
            color = entry['color']
            if color not in combined_data:
                combined_data[color] = {'color': color, 'count': 0, 'weight': 0}
            combined_data[color]['count'] += entry['count']
            combined_data[color]['weight'] += entry['weight']
    return list(combined_data.values())

result = df.groupby('area').apply(combine_data_info).reset_index(name='data_info')

result.to_csv('result.txt', sep='\t', index=False)

字符串

输出：

area    data_info
NY  [{'color': 'Yellow', 'count': 3, 'weight': 5}, {'color': 'Blue', 'count': 7, 'weight': 17}, {'color': 'Red', 'count': 7, 'weight': 3}, {'color': 'Black', 'count': 22, 'weight': 165}]
SF  [{'color': 'Green', 'count': 1, 'weight': 14}, {'color': 'Yellow', 'count': 9, 'weight': 2}]

型

使用自定义函数和groupby.agg：

from itertools import chain
def agg_dict(s):
    return (pd.DataFrame(chain.from_iterable(s))
              .groupby('color', as_index=False).sum()
              .to_dict('records')
           )

out = df.groupby('area', as_index=False)['data_info'].agg(agg_dict)

字符串
输出量：

area                                                                                                                                                                             data_info
0   NY  [{'color': 'Black', 'count': 2, 'weight': 15}, {'color': 'Blue', 'count': 7, 'weight': 17}, {'color': 'Red', 'count': 7, 'weight': 3}, {'color': 'Yellow', 'count': 3, 'weight': 5}]
1   SF                                                                                          [{'color': 'Green', 'count': 1, 'weight': 14}, {'color': 'Yellow', 'count': 9, 'weight': 2}]

型