python pandas 中的日期时间戳和 datetime64[ns, UTC] 比较

Question

由于使用 pandas DataFrames 的过滤过程的影响，我很困惑。 我正在尝试获取介于某些日期之间的行，但我的结果 DataFrame 为空。 我确信那个时期的数据是存在的。

df.info()返回“opentime”是类型：`opendate 440383 non-null datetime64[ns, UTC]

代码片段：

from datetime import timedelta
from datetime import datetime

current_date = pd.datetime.now()
t_delta_week = timedelta(days=7)
t_delta_year = timedelta(days=365)

#CurrentDate
date_start2020 = pd.Timestamp(current_date - t_delta_week, unit='ms')
date_end2020 = pd.Timestamp(current_date, unit='ms')

date_start2020 = date_start2020.tz_localize('utc')
date_end2020 = date_end2020.tz_localize('utc')


#LastYearDate
date_start2019 = pd.Timestamp(current_date - t_delta_year - t_delta_week, unit='ms')
date_end2019 = pd.Timestamp(current_date - t_delta_year, unit='ms')

date_start2019 = date_start2019.tz_localize('utc')
date_end2019 = date_end2019.tz_localize('utc')


df2020_2019['opendate'] = pd.to_datetime(df2020_2019['opendate'], unit='ms') 
mask = (df2020_2019['opendate'] > date_start2020) & (df2020_2019['opendate'] <= date_end2020)
df_currYear = df2020_2019.loc[mask]

df_currYear

返回的 DataFrame 为空

感谢帮助： ：）

编辑：

也许这可能会有所帮助：“opendate”是生成列并使用以下代码片段创建：

import pandas as pd
fmt = '%Y-%m-%dT%H:%M:%S'
df2020_2019.dropna(subset=['opentime_TS'], inplace=True)
df2020_2019['opendate'] = pd.to_datetime(df2020_2019['opentime_TS'], utc=True, format=fmt, errors='ignore')

此外，我还放了一些数据样本的head()打印件。 由于隐私，我无法提供 df 的记录:)

Answer 1

好吧，我的错。 由于专注于tz TypeErrors，我刚刚变得盲目......我采用了已经过时的错误数据源：）适用于正确数据的最终解决方案：

from datetime import timedelta
from datetime import datetime
import pandas as pd

fmt = '%Y-%m-%dT%H:%M:%S'
df2020_2019.dropna(subset=['opentime_TS'], inplace=True)
df2020_2019['opendate'] = pd.to_datetime(df2020_2019['opentime_TS'], utc=True, format=fmt, errors='ignore')
df2020_2019.info()

current_date = pd.Timestamp.now()
t_delta_week = timedelta(days=7)
t_delta_year = timedelta(days=365)

#CurrentDate
date_start2020 = pd.Timestamp(current_date - t_delta_week, unit='ms')
date_end2020 = pd.Timestamp(current_date, unit='ms')

date_start2020 = date_start2020.tz_localize('utc')
date_end2020 = date_end2020.tz_localize('utc')


#LastYearDate
date_start2019 = pd.Timestamp(current_date - t_delta_year - t_delta_week, unit='ms')
date_end2019 = pd.Timestamp(current_date - t_delta_year, unit='ms')

date_start2019 = date_start2019.tz_localize('utc')
date_end2019 = date_end2019.tz_localize('utc')

df2020_2019['opendate'] = pd.to_datetime(df2020_2019['opendate'], unit='ms') 

df_currYear = df2020_2019[df2020_2019["opendate"] > date_start2020]
df_lastYear = df2020_2019[df2020_2019["opendate"].between(date_start2019, date_end2019)]

df_currYear