Pandas 日期时间转换 CET/CEST 到 UTC

Question

我有一个 df 日期时间列，我想将其从 Europe/Copenhgaen t.z 转换为 UTC，但我只是不断在 UTC 列中收到重复的条目。发生这种情况的原因是我制作日期时间列的方式。

我的数据以 df 形式出现，其中包含日期（开始时间设置为 00:00）、位置、分辨率、ID 和数量。对于给定日期，每个 ID 有 23 到 25 个数字位置，具体取决于日期，为了简单起见，我们假设分辨率只是每小时，并且我们有一个 ID。

我有一个函数，可以将日期、分辨率和位置转换为日期时间，然后将其设置为所需的时区，然后将其转换为 UTC。但是，我遇到的问题是 UTC 日期列包含重复的日期而不是所需的日期，2024-10-27 01:00

示例：

import pytz
import pandas as pd
import numpy as np
from datetime import datetime, timedelta

# Set up date range from 25th October 2024 to 28th October 2024
date_range = pd.date_range(start="2024-10-25", end="2024-10-28", freq="D")

# Initialize an empty list to collect data
data = []

# Populate data for each date
for date in date_range:
    # Set number of positions based on date
    if date == datetime(2024, 10, 27):
        positions = range(1, 26)  # 25 positions on 27th October
    else:
        positions = range(1, 25)  # 24 positions on other dates
    
    # Generate data entries for each position
    for pos in positions:
        data.append({
            "Date": date,
            "position": pos,
            "ID": "A",
            "quantity": np.random.randint(1, 100),
            "resolution": 'PT1H'
        })

# Create DataFrame
df = pd.DataFrame(data)

copenhagen_tz = pytz.timezone('Europe/Copenhagen')
df['Date'] = pd.to_datetime(df['Date'],utc=False)

# Define a dictionary to map resolution text to timedelta values
resolution_map = {
    'PT1H': timedelta(hours=1),
    'PT30M': timedelta(minutes=30),
    'PT15M': timedelta(minutes=15),
    'PT5M': timedelta(minutes=5)
}

# Calculate the exact datetime by applying the resolution offset
def calculate_exact_date(row):
    base_time = row['Date']
    offset = (row['position'] - 1) * resolution_map[row['resolution']]
    # Handle DST ambiguity by using `ambiguous=True` to interpret ambiguous times consistently
    exact_date = base_time + offset

    return copenhagen_tz.normalize(exact_date)

# Convert 'Date' to datetime in UTC, then localize to Europe/Copenhagen
df['Date'] = pd.to_datetime(df['Date']).dt.tz_localize('CET')

df['Date_Local'] = df.apply(calculate_exact_date, axis=1)
df['Date_UTC'] = pd.to_datetime(df['Date_Local']).dt.tz_convert('UTC') 

我有什么想法可以以不同的方式做到这一点吗？我无法更改收到的数据的格式，它始终有一个位置和开始日期时间。

Answer 1

由于CEST到CET时间变化，有两个0200时间。 如果您以当地时间进行时间计算，您最终会得到两个相同的 UTC 时间。 以 UTC 时间计算：

import pandas as pd
from datetime import datetime, timedelta
from zoneinfo import ZoneInfo

zone = ZoneInfo('Europe/Copenhagen')
date_range = pd.date_range(start='2024-10-27', end='2024-10-27', freq='D', tz=zone)

data = []
for date in date_range:
    for pos in range(1, 26):
        data.append({'Date': date, 'position': pos})

df = pd.DataFrame(data)

def calculate_exact_date(row):
    base_time = row['Date']
    offset = (row['position'] - 1) * timedelta(hours=1)
    exact_date = base_time + offset
    return exact_date.astimezone(zone)

df['Date'] = pd.to_datetime(df['Date']).dt.tz_convert('UTC')
df['Date_Local'] = df.apply(calculate_exact_date, axis=1)
df['Date_UTC'] = pd.to_datetime(df['Date_Local']).dt.tz_convert('UTC')
del df['Date']  # To see Date_Local and Date_UTC in the default print below
print(df)

输出：

    position                Date_Local                  Date_UTC
0          1 2024-10-27 00:00:00+02:00 2024-10-26 22:00:00+00:00
1          2 2024-10-27 01:00:00+02:00 2024-10-26 23:00:00+00:00
2          3 2024-10-27 02:00:00+02:00 2024-10-27 00:00:00+00:00
3          4 2024-10-27 02:00:00+01:00 2024-10-27 01:00:00+00:00
4          5 2024-10-27 03:00:00+01:00 2024-10-27 02:00:00+00:00
5          6 2024-10-27 04:00:00+01:00 2024-10-27 03:00:00+00:00
6          7 2024-10-27 05:00:00+01:00 2024-10-27 04:00:00+00:00
7          8 2024-10-27 06:00:00+01:00 2024-10-27 05:00:00+00:00
8          9 2024-10-27 07:00:00+01:00 2024-10-27 06:00:00+00:00
9         10 2024-10-27 08:00:00+01:00 2024-10-27 07:00:00+00:00
10        11 2024-10-27 09:00:00+01:00 2024-10-27 08:00:00+00:00
11        12 2024-10-27 10:00:00+01:00 2024-10-27 09:00:00+00:00
12        13 2024-10-27 11:00:00+01:00 2024-10-27 10:00:00+00:00
13        14 2024-10-27 12:00:00+01:00 2024-10-27 11:00:00+00:00
14        15 2024-10-27 13:00:00+01:00 2024-10-27 12:00:00+00:00
15        16 2024-10-27 14:00:00+01:00 2024-10-27 13:00:00+00:00
16        17 2024-10-27 15:00:00+01:00 2024-10-27 14:00:00+00:00
17        18 2024-10-27 16:00:00+01:00 2024-10-27 15:00:00+00:00
18        19 2024-10-27 17:00:00+01:00 2024-10-27 16:00:00+00:00
19        20 2024-10-27 18:00:00+01:00 2024-10-27 17:00:00+00:00
20        21 2024-10-27 19:00:00+01:00 2024-10-27 18:00:00+00:00
21        22 2024-10-27 20:00:00+01:00 2024-10-27 19:00:00+00:00
22        23 2024-10-27 21:00:00+01:00 2024-10-27 20:00:00+00:00
23        24 2024-10-27 22:00:00+01:00 2024-10-27 21:00:00+00:00
24        25 2024-10-27 23:00:00+01:00 2024-10-27 22:00:00+00:00