python : group by columns with columns values that are grouped by occurs only once and retain all other columns

Question

假设我有一个包含以下列的字典

dict_ = [
    {'key1': 'value1',
     'key2': 'value2',
     'key3': 'value3',
     'key4': 'value4',
     'key5': 'value5',
     'nested_dicts' : [
         {'nested_key1' : 'nested_val_1',
          'nested_key2' : 'nested_val_2',
          'nested_key3' : 'nested_val_3',
          'nested_key4' : 'nested_val_4'
          },
         {'nested_key1' : 'nested_val_5',
          'nested_key2' : 'nested_val_6',
          'nested_key3' : 'nested_val_7',
          'nested_key4' : 'nested_val_8'
        }
    ]},
    {
     'key1': 'value6',
     'key2': 'value7',
     'key3': 'value8',
     'key4': 'value9',
     'key5': 'value10',
     'nested_dicts' : [
         {'nested_key1' : 'nested_val_9',
          'nested_key2' : 'nested_val_10',
          'nested_key3' : 'nested_val_11',
          'nested_key4' : 'nested_val_12'
          },
         {'nested_key1' : 'nested_val_9',
          'nested_key2' : 'nested_val_10',
          'nested_key3' : 'nested_val_11',
          'nested_key4' : 'nested_val_12'
        }
    ]}
]

我需要按值分组，这样被分组的列只出现一次，所有其他列按原样显示。 预期输出是这样的:

 key1    key2    key3    key4    key5    nested_key1   nested_key2     nested_key3     nested_key4
value1  value2  value3  value4  value5   nested_val_1  nested_val_2    nested_val_3    nested_val_4
                                         nested_val_5  nested_val_6    nested_val_7    nested_val_8
value6  value7  value8  value9  value10  nested_val_9  nested_val_10   nested_val_11   nested_val_12
                                         nested_val_13  nested_val_14  nested_val_15   nested_val_16

任何使用 groupby 或 multiIndex 或其他 pandas 函数的解决方案都将被接受。

Answer 1

这是一个可能的解决方案:

from pandas import json_normalize

df = json_normalize(
    dict_,
    record_path=['nested_dicts'],
    meta = ['key1', 'key2', 'key3', 'key4', 'key5']
)
df = df.set_index(['key1', 'key2', 'key3', 'key4', 'key5'])

这里我使用了一个json_normalize函数来扁平化字典，然后我将索引设置为key1-key5。