将字典转换为多索引数据框

Question

我有一个构造如下的默认字典：

data = defaultdict(dict)
symbol_list = [
    'ETHUSDT',
    'BTCUSDT'
]
for symbol in symbol_list:
    data[symbol] = load_binance_data(c, symbol, '2021-12-23', timeframe='5m')

这是作为值存储在字典中的数据框的轴：

[DatetimeIndex(['2021-12-23 00:05:00', '2021-12-23 00:10:00',
               '2021-12-23 00:15:00', '2021-12-23 00:20:00',
               '2021-12-23 00:25:00', '2021-12-23 00:30:00',
               '2021-12-23 00:35:00', '2021-12-23 00:40:00',
               '2021-12-23 00:45:00', '2021-12-23 00:50:00',
               ...
               '2021-12-24 19:05:00', '2021-12-24 19:10:00',
               '2021-12-24 19:15:00', '2021-12-24 19:20:00',
               '2021-12-24 19:25:00', '2021-12-24 19:30:00',
               '2021-12-24 19:35:00', '2021-12-24 19:40:00',
               '2021-12-24 19:45:00', '2021-12-24 19:50:00'],
              dtype='datetime64[ns]', name='time', length=526, freq=None), Index(['open', 'high', 'low', 'close', 'volume'],
      dtype='object')]

我想将此字典转换为具有多个索引的单个数据框，如下所示：

[DatetimeIndex(['2021-12-23 00:05:00', '2021-12-23 00:10:00',
                   '2021-12-23 00:15:00', '2021-12-23 00:20:00',
                   '2021-12-23 00:25:00', '2021-12-23 00:30:00',
                   '2021-12-23 00:35:00', '2021-12-23 00:40:00',
                   '2021-12-23 00:45:00', '2021-12-23 00:50:00',
                   ...
                   '2021-12-24 19:05:00', '2021-12-24 19:10:00',
                   '2021-12-24 19:15:00', '2021-12-24 19:20:00',
                   '2021-12-24 19:25:00', '2021-12-24 19:30:00',
                   '2021-12-24 19:35:00', '2021-12-24 19:40:00',
                   '2021-12-24 19:45:00', '2021-12-24 19:50:00'],
              dtype='datetime64[ns]', name='time', freq=None), 
              MultiIndex([
                  ('open', 'ETHUSDT'),
                  ('open', 'BTCUSDT'),
                  ('high', 'ETHUSDT'),
                  ('high', 'BTCUSDT'),
                  ('low', 'ETHUSDT'),
                  ('low', 'BTCUSDT'),
                  ('close', 'ETHUSDT'),
                  ('close', 'BTCUSDT'),
                  ('volume', 'ETHUSDT'),
                  ('volume', 'BTCUSDT')],
           names=['Attributes', 'Symbols'])]

我该如何进行这种转换？

提前致谢，

Answer 1

如果我理解正确，你有两个 DataFrame：

• DataFrame 1，我们称之为SYMBOL1：

  	open	high	low	close	volume
  time
  2021-12-23 00:05:00	1	3	5	7	9
  2021-12-23 00:10:00	2	4	6	8	10

• 还有DataFrame 2，我们称之为SYMBOL2：

  	open	high	low	close	volume
  time
  2021-12-23 00:05:00	-1	-3	-5	-7	-9
  2021-12-23 00:10:00	-2	-4	-6	-8	-10

您正试图将其转换为具有上述形状轴的 DataFrame。如果是这样，那么这是一种方法：

import pandas as pd

# Code to Create the DataFrames in the example :
d = {'open': [1, 2], 'high': [3, 4], 'low':[5,6], 'close':[7,8], 'volume':[9,10]}
df1 = pd.DataFrame(d, index=pd.DatetimeIndex(
    ['2021-12-23 00:05:00', '2021-12-23 00:10:00'], name='time'))
df2 = df1*-1

# The transformation : 
new_axis_df1 = pd.MultiIndex.from_product(
    [df1.axes[1].values, ['SYMBOL1']], names=['Attributes', 'Symbols'])
new_df1 = df1.set_axis(new_axis_df1, axis='columns')

new_axis_df2 = pd.MultiIndex.from_product(
    [df2.axes[1].values, ['SYMBOL2']], names=['Attributes', 'Symbols'])
new_df2 = df2.set_axis(new_axis_df2, axis='columns')

# Merging the transformed DataFrames
final_df = new_df1.merge(new_df2, on='time')

#Result : 
print(final_df.axes)

这会产生并输出以下 DataFrame 轴：

[DatetimeIndex(['2021-12-23 00:05:00', '2021-12-23 00:10:00'], dtype='datetime64[ns]', name='time', freq=None),
 MultiIndex([(  'open', 'SYMBOL1'),
             (  'high', 'SYMBOL1'),
             (   'low', 'SYMBOL1'),
             ( 'close', 'SYMBOL1'),
             ('volume', 'SYMBOL1'),
             (  'open', 'SYMBOL2'),
             (  'high', 'SYMBOL2'),
             (   'low', 'SYMBOL2'),
             ( 'close', 'SYMBOL2'),
             ('volume', 'SYMBOL2')],
            names=['Attributes', 'Symbols'])]

final_df 的示例如下所示：

Attributes	open	high	low	close	volume	open	high	low	close	volume
Symbols	SYMBOL1	SYMBOL1	SYMBOL1	SYMBOL1	SYMBOL1	SYMBOL2	SYMBOL2	SYMBOL2	SYMBOL2	SYMBOL2
time
2021-12-23 00:05:00	1	3	5	7	9	-1	-3	-5	-7	-9
2021-12-23 00:10:00	2	4	6	8	10	-2	-4	-6	-8	-10