pandas - dtype=datetime64[ns, UTC] 和时间戳之间的无效比较

Question

这是数据集Dataset的一部分

1. Ordered Quantity Time Stamp Medicine Name Actual Price
   14162 3 2020-10-17 10:06:33.790000+00:00 27699 263.82
   15747 2 2020-10-28 09:14:46.690000+00:00 27699 175.88
   24866 2 2020-12-27 14:16:26.259000+00:00 27699 175.88
   25962 2 2021-01-03 18:35:04.337000+00:00 27699 175.88
   29188 2 2021-01-28 08:17:40.975000+00:00 27699 175.88
   29565 2 2021-02-02 04:15:58.496000+00:00 27699 175.88
   30863 2 2021-02-09 07:19:44.897000+00:00 27699 175.88
   33931 2 2021-03-02 03:39:52.087000+00:00 27699 175.88
   35537 2 2021-03-13 15:19:58.131000+00:00 27699 175.88
   37429 1 2021-03-27 05:15:31.772000+00:00 27699 87.94
   47099 1 2021-05-03 07:16:10.483000+00:00 27699 87.94
   49135 2 2021-05-09 08:03:35.300000+00:00 27699 175.88
   66057 1 2021-07-07 07:50:21.414000+00:00 27699 87.94
   67937 1 2021-07-13 06:44:20.875000+00:00 27699 87.94

`

  import pandas as pd
  data = pd.read_csv("data.csv")
  data["Time Stamp"] = pd.to_datetime(data["Time Stamp"])
  t1 = pd.Timestamp("2020-06-01 00:00:00.054000+00:00")
  t2 = t1.date() + pd.DateOffset(months = 6)
  data[(data["Time Stamp"] > t1) & (data["Time Stamp"] < t2)]
`

TypeError: dtype=datetime64[ns, UTC] 和时间戳之间的比较无效

为什么会出现此错误？

Answer 1

尝试:

在比较中不要使用 t2，而是使用 t2.tz_localize('utc'):

data[(data["Time Stamp"] > t1) & (data["Time Stamp"] < t2.tz_localize('utc'))]

或

使用normalize() 方法代替date() 方法:

t2=t1.normalize() + pd.DateOffset(months = 6)