将列表列表（嵌套列表）转换为数据框

Question

我有一个嵌套列表，其中每个 'nested list' 按 10 分钟间隔分组（使用 timedelta 函数）。

print(groups)
-----------------------results---------------------------
[[[1, Timestamp('2019-01-22 00:54:27')]],
 [[2, Timestamp('2019-01-22 08:37:04')]],
 [[3, Timestamp('2019-01-22 10:57:40')],
  [4, Timestamp('2019-01-22 10:57:43')]],
 [[5, Timestamp('2019-01-22 11:09:07')],
  [6, Timestamp('2019-01-22 11:16:18')],
  [7, Timestamp('2019-01-22 11:16:23')],
  [8, Timestamp('2019-01-22 11:16:25')]],
 [[9, Timestamp('2019-01-22 11:35:03')],
  [10, Timestamp('2019-01-22 11:35:35')]]...

这些信息最初来自完整的数据框。我正在尝试将此嵌套列表转换为数据框格式，并带有一个名为“Group”的单独列。这意味着第一个 'nested list' 将是第 1 组，第二个 'nested list' 将是第 2 组。

正在考虑在转换为数据框之前先转换为字典。但是，它们也不起作用。

{k:row[0] for row in groups for k in row[1:]}
dict((k[0], k[1:]) for k in groups)
error: unhashable type: 'list'

总而言之，我希望编写一个函数来自动化我的流程，如下所示：

df_1 = pd.DataFrame(groups[1],columns=['index','datetime'])
df_1['group']=1

df_2 = pd.DataFrame(groups[2],columns=['index','datetime'])
df_2['group']=2

df_3 = pd.DataFrame(groups[3],columns=['index','datetime'])
df_3['group']=3
... etc...

很想听听你们是如何处理这个问题的！谢谢

Answer 1

这是您要实现的目标吗？

Try it online!

import pandas as pd
from pandas import Timestamp

l = [
 [[1, Timestamp('2019-01-22 00:54:27')]],
 [[2, Timestamp('2019-01-22 08:37:04')]],
 [[3, Timestamp('2019-01-22 10:57:40')],
  [4, Timestamp('2019-01-22 10:57:43')]],
 [[5, Timestamp('2019-01-22 11:09:07')],
  [6, Timestamp('2019-01-22 11:16:18')],
  [7, Timestamp('2019-01-22 11:16:23')],
  [8, Timestamp('2019-01-22 11:16:25')]],
 [[9, Timestamp('2019-01-22 11:35:03')],
  [10, Timestamp('2019-01-22 11:35:35')]],
]

df = pd.DataFrame(
    [e1 + [g] for g, e0 in enumerate(l) for e1 in e0],
    columns = ['id', 'time', 'group'],
)

print('As one DataFrame:\n', df)

dfs = [
    pd.DataFrame(
        [e1 + [g] for e1 in e0],
        columns = ['id', 'time', 'group'],
    )
    for g, e0 in enumerate(l)
]

print('\nAs separate DataFrames:')

for df in dfs:
    print('---------------------')
    print(df)

产出

As one DataFrame:
    id                time  group
0   1 2019-01-22 00:54:27      0
1   2 2019-01-22 08:37:04      1
2   3 2019-01-22 10:57:40      2
3   4 2019-01-22 10:57:43      2
4   5 2019-01-22 11:09:07      3
5   6 2019-01-22 11:16:18      3
6   7 2019-01-22 11:16:23      3
7   8 2019-01-22 11:16:25      3
8   9 2019-01-22 11:35:03      4
9  10 2019-01-22 11:35:35      4

As separate DataFrames:
---------------------
   id                time  group
0   1 2019-01-22 00:54:27      0
---------------------
   id                time  group
0   2 2019-01-22 08:37:04      1
---------------------
   id                time  group
0   3 2019-01-22 10:57:40      2
1   4 2019-01-22 10:57:43      2
---------------------
   id                time  group
0   5 2019-01-22 11:09:07      3
1   6 2019-01-22 11:16:18      3
2   7 2019-01-22 11:16:23      3
3   8 2019-01-22 11:16:25      3
---------------------
   id                time  group
0   9 2019-01-22 11:35:03      4
1  10 2019-01-22 11:35:35      4

将列表列表（嵌套列表）转换为数据框

Converting List of List (nested list) into Dataframe

python

nested

timedelta

dataframe

pandas