为什么 datetime 可以处理正时区 (UTC) 而不是负时区 (UTC-)？

Question

我一直在使用 Pandas 将 .CSV 文件转换为可在另一个系统上读取的格式，并且我即将完成它，但我无法让它与负时区 (UTC- 1、-2 等）

这是我正在使用的代码，它不是最整洁的，但它可以完成 UTC+ 时区的工作，你知道为什么它可能无法正确处理 UTC- 时区吗？

import pandas as pd
from datetime import datetime
pd.set_option('display.max_rows', 500)
pd.set_option('display.max_columns', 500)
pd.set_option('display.width', 1000)

import sys
import io
sys.stdout = io.TextIOWrapper(sys.stdout.detach(), encoding = 'utf-8')
sys.stderr = io.TextIOWrapper(sys.stderr.detach(), encoding = 'utf-8')

df = pd.read_csv('CONCACAF_First_Round.csv', index_col=False)
df['starttime'] = df['starttime'].str.replace('\s+', '')
df.insert(loc=2, column='season', value='2020')
df.insert(loc=6, column='awayscore', value='')
df.insert(loc=8, column='round_a', value='1')
df['venue'] = df['venue'].str.split(',').str[0]
df[['homescore', 'awayscore']] = df['homescore'].str.split('–',expand=True)
df['awayscore'] = df['awayscore'].str.split(' ').str[0]
df['starttime'] = df['starttime'].str.replace('UTC', ' UTC')
df['datepicker'] = df['datepicker'] + (' ') + df['starttime']
del df['starttime']
df[['datepicker', 'time', 'UTC']] = df.datepicker.str.split(" ", expand=True)
df['datepicker'] = df['datepicker'] + ' ' + df['time']
del df['time']
df['datepicker'] = df['datepicker'] + ' ' + df['UTC']
df['datepicker'] = df['datepicker'].str.replace('±', '+')
df['datepicker'] = df['datepicker'].str.replace('UTC', '')
del df['UTC']
df['datepicker'] = pd.to_datetime(df['datepicker'], utc=True)
df.insert(loc=1, column='starttime', value='')
df['starttime'] = df['datepicker'].dt.strftime('%H:%M:%S')
df['datepicker'] = df['datepicker'].dt.strftime('%Y-%m-%d')
print(df.head(10))

这是它返回的负时区

dateutil.parser._parser.ParserError: Unknown string format: 2015-03-25 19:30 −4

这是它在 UTC+ 时区正常工作时返回的内容

datepicker starttime season  hometeam awayteam homescore awayscore              venue round_a
0  2019-09-04  13:00:00   2020  Ethiopia  Lesotho         0         0  Bahir Dar Stadium       1

数据：CONCACAF_First_Round.csv

# copy the data to the clipboard and read with
df = pd.read_clipboard(sep=',')

datepicker,starttime,hometeam,awayteam,homescore,venue
2015-03-25,19:30 UTC−4,Bahamas,Bermuda,0–5,"Thomas Robinson Stadium, Nassau"
2015-03-29,15:00 UTC−3,Bermuda,Bahamas,3–0,"Bermuda National Stadium, Devonshire"
2015-03-26,19:00 UTC−4,British Virgin Islands,Dominica,2–3,"Windsor Park, Roseau (Dominica)[note 2]"
2015-03-29,17:00 UTC−4,Dominica,British Virgin Islands,0–0,"Windsor Park, Roseau"
2015-03-22,19:00 UTC−4,Barbados,U.S. Virgin Islands,0–1,"Barbados National Stadium, Bridgetown"
2015-03-26,15:30 UTC−4,U.S. Virgin Islands,Barbados,0–4,"Addelita Cancryn Junior High School Ground, Charlotte Amalie"
2015-03-23,20:00 UTC−4,Saint Kitts and Nevis,Turks and Caicos Islands,6–2,"Warner Park, Basseterre"
2015-03-26,19:00 UTC−4,Turks and Caicos Islands,Saint Kitts and Nevis,2–6,"TCIFA National Academy, Providenciales"
2015-03-23,18:00 UTC−6,Nicaragua,Anguilla,5–0,"Nicaragua National Football Stadium, Managua"
2015-03-29,17:00 UTC−4,Anguilla,Nicaragua,0–3,"Ronald Webster Park, The Valley"
2015-03-25,20:00 UTC−6,Belize,Cayman Islands,0–0,"FFB Stadium, Belmopan"
2015-03-29,19:00 UTC−5,Cayman Islands,Belize,1–1,"Truman Bodden Sports Complex, George Town"
2015-03-27,20:00 UTC−4,Curaçao,Montserrat,2–1,"Ergilio Hato Stadium, Willemstad"
2015-03-31,19:00 UTC−4,Montserrat,Curaçao,2–2,"Blakes Estate Stadium, St. John's"

Answer 1

您的“减号”并不都是减号。例如在您的错误消息中：

Unknown string format: 2015-03-25 19:30 −4

如果您将看起来像减号的倒数第二个字符作为字符串复制到 Python 中，您将看到：

>>> '−'
'\xe2\x88\x92'

那是 U+2212，一个特殊的Unicode "minus sign"，它与键盘上的“-”不同。

您需要将那些 Unicode“减号”替换为纯 ASCII 减号，然后它应该可以工作。

有一个Python-ideas discussion about this in 2013，“Unicode 减号在数字转换中”，其中one person said：

就支持非 ASCII 加号和减号而言，我原则上很热衷，但在实践中不冷不热。我认为这将是一个很好的选择，如果有人确定应该接受哪些字符，我会支持将其添加为新功能。但我不认为缺乏对非 ASCII 数字符号的支持是一个错误。

这个话题最终没有结果，因为每个人都多次将他们的爱好马打死，还带着诸如泰国数字是否应该在float() 中使用这样的附带话题。如果有人接手这项任务，更专注于支持 U+2212 和一些 Unicode“加号”符号可能会更好。

Answer 2

fix_starttime击穿

• 数据的主要问题是− 不是正确的符号-。请注意，它会稍长一些。
  • 将'−' 替换为'-'
  • 将' UTC' 替换为':00 '
  • 将'19:30 UTC−4' 转换为'19:30:00 -4'
• fix_date 函数使用'19:30:00 -4' 执行以下操作
  • y = v.split(' ') → y = ['19:30:00', '-4']
  • z= y[1][0] + y[1][1:].zfill(2) + '00' → z = '-' + '0400' → z = '-0400'
    • str.zfill 用零填充 str 以适应长度 width，在这种情况下为 2。所以'4' 变为'04' 而'10' 保持'10'。
  • t = y[0] → t = '19:30:00'
  • return 19:30:00-0400
• 将datepicker 和格式正确的starttime 组合在一起，并将其转换为datetime64[ns, UTC] 格式的列。

import pandas as pd

# read the file
df = pd.read_csv('CONCACAF_First_Round.csv')

# function
def fix_starttime(v: str) -> str:
    y = v.split(' ')
    z = y[1][0] + y[1][1:].zfill(2) + '00'
    t = y[0]
    return t + z


# fix starttime
df['starttime'] = df.starttime.str.replace('−', '-').str.replace(' UTC', ':00 ').apply(fix_starttime)

# create datetime column in datetime64[ns, UTC] format
df['utc_datetime'] = pd.to_datetime(df.datepicker + 'T' + df.starttime, utc=True)

合并

# read the file
df = pd.read_csv('CONCACAF_First_Round.csv')

# fix starttime
df['starttime'] = df.starttime.str.replace('−', '-').str.replace(' UTC', ':00 ').str.split(' ').apply(lambda y: y[0] + y[1][0] + y[1][1:].zfill(2) + '00')

# create datetime column in datetime64[ns, UTC] format
df['utc_datetime'] = pd.to_datetime(df.datepicker + 'T' + df.starttime, utc=True)

附加功能

• 一旦有日期时间格式的列，使用 pandas .dt accessor 提取日期时间组件。

# extract time from utc_datetime
df.utc_datetime.dt.time

# extract date from utc_datetime
df.utc_datetime.dt.date

结果

 datepicker      starttime                  hometeam                  awayteam homescore                                                         venue              utc_datetime
 2015-03-25  19:30:00-0400                   Bahamas                   Bermuda       0–5                               Thomas Robinson Stadium, Nassau 2015-03-25 23:30:00+00:00
 2015-03-29  15:00:00-0300                   Bermuda                   Bahamas       3–0                          Bermuda National Stadium, Devonshire 2015-03-29 18:00:00+00:00
 2015-03-26  19:00:00-0400    British Virgin Islands                  Dominica       2–3                       Windsor Park, Roseau (Dominica)[note 2] 2015-03-26 23:00:00+00:00
 2015-03-29  17:00:00-0400                  Dominica    British Virgin Islands       0–0                                          Windsor Park, Roseau 2015-03-29 21:00:00+00:00
 2015-03-22  19:00:00-0400                  Barbados       U.S. Virgin Islands       0–1                         Barbados National Stadium, Bridgetown 2015-03-22 23:00:00+00:00
 2015-03-26  15:30:00-0400       U.S. Virgin Islands                  Barbados       0–4  Addelita Cancryn Junior High School Ground, Charlotte Amalie 2015-03-26 19:30:00+00:00
 2015-03-23  20:00:00-0400     Saint Kitts and Nevis  Turks and Caicos Islands       6–2                                       Warner Park, Basseterre 2015-03-24 00:00:00+00:00
 2015-03-26  19:00:00-0400  Turks and Caicos Islands     Saint Kitts and Nevis       2–6                        TCIFA National Academy, Providenciales 2015-03-26 23:00:00+00:00
 2015-03-23  18:00:00-0600                 Nicaragua                  Anguilla       5–0                  Nicaragua National Football Stadium, Managua 2015-03-24 00:00:00+00:00
 2015-03-29  17:00:00-0400                  Anguilla                 Nicaragua       0–3                               Ronald Webster Park, The Valley 2015-03-29 21:00:00+00:00
 2015-03-25  20:00:00-0600                    Belize            Cayman Islands       0–0                                         FFB Stadium, Belmopan 2015-03-26 02:00:00+00:00
 2015-03-29  19:00:00-0500            Cayman Islands                    Belize       1–1                     Truman Bodden Sports Complex, George Town 2015-03-30 00:00:00+00:00
 2015-03-27  20:00:00-0400                   Curaçao                Montserrat       2–1                              Ergilio Hato Stadium, Willemstad 2015-03-28 00:00:00+00:00
 2015-03-31  19:00:00-0400                Montserrat                   Curaçao       2–2                             Blakes Estate Stadium, St. John's 2015-03-31 23:00:00+00:00