如何按行业而不是按时间分组

Question

                                   open     high     low       close    Volume
ctime                                                   
2022-11-07 01:00:00.012000+01:00  20900.0  20900.0  20900.0  20900.0     209.00
2022-11-07 01:00:00.019000+01:00  20900.1  20900.1  20900.1  20900.1    1254.00
2022-11-07 01:00:00.111000+01:00  20900.0  20900.0  20900.0  20900.0   11390.50
2022-11-07 01:00:00.188000+01:00  20900.1  20900.1  20900.1  20900.1      20.90
2022-11-07 01:00:00.194000+01:00  20900.0  20900.0  20900.0  20900.0    4995.10
2022-11-07 01:00:00.203000+01:00  20900.0  20900.0  20900.0  20900.0     209.00
2022-11-07 01:00:00.217000+01:00  20900.0  20900.0  20900.0  20900.0  239827.50
2022-11-07 01:00:00.237000+01:00  20900.0  20900.0  20900.0  20900.0      20.90
2022-11-07 01:00:03.028000+01:00  20900.0  20900.0  20900.0  20900.0      20.90
2022-11-07 01:00:03.743000+01:00  20900.0  20900.0  20900.0  20900.0      41.80

我需要集中在五个交易的时期，也就是说，开盘是第一行，收盘是第五笔交易，最低和最高是对应的最小和最高的5笔交易，同样是成交量。一旦转换完成，公开的示例中应该只剩下2行。

我试过寻找熊猫的近似值，但我没有找到令人满意的解决方案。

Answer 1

定义一个新的系列，将交易分成5组，然后按组汇总：

trade_group = np.arange(len(df)) // 5

# The `open` of the group is the `open` of the first trade in the group
# `high` is max of the `high` in the group
# and so on...
df.groupby(trade_group).agg({
    "open": "first",
    "high": "max",
    "low": "min",
    "close": "last",
    "Volume": "sum"
})