Arrow是一个Python库，它提供了一种合理且对人类友好的方法来创建、操作、格式化和转换日期、时间和时间戳。它实现了对datetime类型的更新，填补了功能上的空白，提供了一个智能的模块API，支持许多常见的创建场景。简单来说，它可以帮助您使用更少的导入和更少的代码来处理日期和时间。
Arrow以时间之箭命名，并且受到moment.js和requests的极大启发。
Python的标准库和其他一些低级模块拥有近乎完整的日期、时间和时区功能，但从可用性的角度来看，它们的工作效果并不理想：
模块太多：datetime, time, calendar, dateutil, pytz等
类型太多：date, time, datetime, tzinfo, timedelta, relativedelta等
时区和时间戳转换冗长且不愉快
通常情况下都是时区不敏感的
功能上存在gap：ISO 8601支持、时区间隔支持、不够人性化
完全实现的，可替代datetime
支持Python 3.6+
默认支持时区感知和UTC
为许多常见输入场景提供超简单的创建选项
shift方法支持相对偏移，包括周
自动格式化和解析字符串
广泛支持ISO 8601标准
时区转换
支持dateutil, pytz和ZoneInfo tzinfo对象
生成时间段、范围、下限和上限，适用于从微秒到年的时间框架
使用不断增长的贡献语言环境人性化日期和时间
可扩展为您自己的Arrow派生类型
全面支持PEP 484风格的类型提示
使用 pip or pipenv:
pip install -U arrow>>> import arrow
>>> arrow.get('2024-06-13T17:23:58.970460+08:00')
<Arrow [2024-06-13T17:23:58.970460+08:00]>
>>> utc = arrow.utcnow()
>>> utc
<Arrow [2024-06-13T08:37:39.950050+00:00]>
>>> utc = utc.shift(hours=-1)
>>> utc
<Arrow [2024-06-13T07:37:39.950050+00:00]>
>>> local = utc.to('Asia/Shanghai')
>>> local
<Arrow [2024-06-13T15:37:39.950050+08:00]>
>>> local.timestamp()
1718264259.95005
>>> local.format()
'2024-06-13 15:37:39+08:00'
>>> local.format('YYYY-MM-DD HH:mm:ss ZZ')
'2024-06-13 15:37:39 +08:00'
>>> local.humanize()
'an hour ago'
>>> local.humanize(locale='zh-cn')
'1小时前'通过now方法创建：
>>> arrow.utcnow()
<Arrow [2024-06-13T09:03:42.591212+00:00]>
>>> arrow.now()
<Arrow [2024-06-13T17:04:00.926319+08:00]>
>>> arrow.now('Asia/Shanghai')
<Arrow [2024-06-13T17:04:25.058905+08:00]>通过时间戳创建 (int or float):
>>> arrow.get(1718264259)
<Arrow [2024-06-13T07:37:39+00:00]>
>>> arrow.get(1718264259.95005)
<Arrow [2024-06-13T07:37:39.950050+00:00]>使用无时区时间、有时区时间，或者灵活指定时区：
>>> arrow.get(datetime.utcnow())
<Arrow [2024-06-13T09:14:15.099808+00:00]>
>>> arrow.get(datetime(2024, 6, 13), 'Asia/Shanghai')
<Arrow [2024-06-13T00:00:00+08:00]>
>>> from dateutil import tz
>>> arrow.get(datetime(2024, 6, 13), tz.gettz('Asia/Shanghai'))
<Arrow [2024-06-13T00:00:00+08:00]>
>>> arrow.get(datetime.now(),tz.gettz('Asia/Shanghai'))
<Arrow [2024-06-13T17:16:02.309002+08:00]>从字符串转换：
>>> arrow.get('2024-06-13 15:30:45', 'YYYY-MM-DD HH:mm:ss')
<Arrow [2024-06-13T15:30:45+00:00]>从文本中查找日期：
>>> arrow.get('June was born in May 1980', 'MMMM YYYY')
<Arrow [1980-05-01T00:00:00+00:00]>自动识别ISO 8601格式的字符串，并解析出相应的日期时间对象，包括时区信息：
>>> arrow.get('2024-06-13T15:34:00.000-08:00')
<Arrow [2024-06-13T15:34:00-08:00]>直接实例化Arrow对象（使用datetime类一样的参数）：
>>> arrow.get(2024, 6, 13)
<Arrow [2024-06-13T00:00:00+00:00]>
>>>
>>> arrow.Arrow(2024, 6, 13)
<Arrow [2024-06-13T00:00:00+00:00]>获取日期时间或时间戳表示形式：
>>> a = arrow.utcnow()
>>> a.datetime
datetime.datetime(2024, 6, 13, 9, 25, 1, 195217, tzinfo=tzutc())获取一个原生的时间，包含tzinfo：
>>> a.naive
datetime.datetime(2024, 6, 13, 9, 25, 1, 195217)
>>> a.tzinfo
tzutc()获取任意日期时间值：
>>> a.year
2024
>>> a.month
6调用datetime类相关函数和属性：
>>> a.date()
datetime.date(2024, 6, 13)
>>> a.time()
datetime.time(9, 25, 1, 195217)获取一个新的Arrow对象，更改其属性，就像处理datetime类型一样：
>>> arw = arrow.utcnow()
>>> arw
<Arrow [2024-06-13T09:33:30.538303+00:00]>
>>> arw.replace(hour=4, minute=40)
<Arrow [2024-06-13T04:40:30.538303+00:00]>或者，将属性向前或向后移动的：
>>> arw.shift(weeks=+3)
<Arrow [2024-07-04T09:33:30.538303+00:00]>
>>> arw.shift(days=-3)
<Arrow [2024-06-10T09:33:30.538303+00:00]>甚至，直接替换时区
>>> arw.replace(tzinfo='US/Pacific')
<Arrow [2024-06-13T09:33:30.538303-07:00]>arrow.utcnow().format('YYYY-MM-DD HH:mm:ss ZZ')
'2024-06-13 09:34:54 +00:00'按名称或tzinfo从UTC转换为其他时区：
>>> utc = arrow.utcnow()
>>> utc
<Arrow [2024-06-13T09:36:22.204239+00:00]>
>>> utc.to('US/Pacific')
<Arrow [2024-06-13T02:36:22.204239-07:00]>
>>> utc.to(tz.gettz('US/Pacific'))
<Arrow [2024-06-13T02:36:22.204239-07:00]>或者使用简写：
>>> utc.to('local')
<Arrow [2024-06-13T17:36:22.204239+08:00]>
>>> utc.to('local').to('utc')
<Arrow [2024-06-13T09:36:22.204239+00:00]>相对于现在时间的人性化表现：
>>> past = arrow.utcnow().shift(hours=-1)
>>> past.humanize()
'an hour ago'或者另一个Arrow对下或日期时间：
>>> present = arrow.utcnow()
>>> future = present.shift(hours=2)
>>> future.humanize(present)
'in 2 hours'将时间表示为相对时间或仅包括时间距离：
>>> present = arrow.utcnow()
>>> future = present.shift(hours=2)
>>> future.humanize(present)
'in 2 hours'
>>> future.humanize(present, only_distance=True)
'2 hours'指定特定的时间粒度（或多个）：
>>> present = arrow.utcnow()
>>> future = present.shift(minutes=66)
>>> future.humanize(present, granularity="minute")
'in 66 minutes'
>>> future.humanize(present, granularity=["hour", "minute"])
'in an hour and 6 minutes'
>>> present.humanize(future, granularity=["hour", "minute"])
'an hour and 6 minutes ago'
>>> future.humanize(present, only_distance=True, granularity=["hour", "minute"])
'an hour and 6 minutes'支持不同时区的语言显示：
>>> future = arrow.utcnow().shift(hours=1)
>>> future.humanize(a, locale='ru')
'через час'
>>> future.humanize(a, locale='zh-cn')
'1小时后'使用人类可读的字符串描述，并使用它转换到过去的时间：
>>> arw = arrow.utcnow()
>>> arw
<Arrow [2024-06-13T09:47:35.616633+00:00]>
>>> earlier = arw.dehumanize("2 days ago")
>>> earlier
<Arrow [2024-06-11T09:47:35.616633+00:00]>或者用它来转换到未来的时间：
>>> arw = arrow.utcnow()
>>> arw
<Arrow [2024-06-13T09:47:35.616633+00:00]>
>>> later = arw.dehumanize("in a month")
>>> later
<Arrow [2024-07-13T09:47:35.616633+00:00]>支持不同时区的语言转换：
>>> later = arw.dehumanize('1小时后', locale="zh-cn")
>>> later
<Arrow [2024-06-13T10:47:35.616633+00:00]>获取任意单位的时间跨度：
>>> arrow.utcnow().span('hour')
(<Arrow [2024-06-13T09:00:00+00:00]>, <Arrow [2024-06-13T09:59:59.999999+00:00]>)
>>> arrow.utcnow().span('day')
(<Arrow [2024-06-13T00:00:00+00:00]>, <Arrow [2024-06-13T23:59:59.999999+00:00]>)或者只获取最大最小值：
>>> arrow.utcnow().floor('hour')
<Arrow [2024-06-13T09:00:00+00:00]>
>>> arrow.utcnow().ceil('hour')
<Arrow [2024-06-13T09:59:59.999999+00:00]>您还可以获取一系列时间跨度：
>>> start = datetime(2024, 6, 13, 12, 30)
>>> end = datetime(2024, 6, 13, 17, 15)
>>> for r in arrow.Arrow.span_range('hour', start, end):
... print(r)
...
(<Arrow [2024-06-13T12:00:00+00:00]>, <Arrow [2024-06-13T12:59:59.999999+00:00]>)
(<Arrow [2024-06-13T13:00:00+00:00]>, <Arrow [2024-06-13T13:59:59.999999+00:00]>)
(<Arrow [2024-06-13T14:00:00+00:00]>, <Arrow [2024-06-13T14:59:59.999999+00:00]>)
(<Arrow [2024-06-13T15:00:00+00:00]>, <Arrow [2024-06-13T15:59:59.999999+00:00]>)
(<Arrow [2024-06-13T16:00:00+00:00]>, <Arrow [2024-06-13T16:59:59.999999+00:00]>)
(<Arrow [2024-06-13T17:00:00+00:00]>, <Arrow [2024-06-13T17:59:59.999999+00:00]>)或者只是获取一系列时间点：
>>> start = datetime(2024, 6, 13, 12, 30)
>>> end = datetime(2024, 6, 13, 17, 15)
>>> for r in arrow.Arrow.range('hour', start, end):
... print(repr(r))
...
<Arrow [2024-06-13T12:30:00+00:00]>
<Arrow [2024-06-13T13:30:00+00:00]>
<Arrow [2024-06-13T14:30:00+00:00]>
<Arrow [2024-06-13T15:30:00+00:00]>
<Arrow [2024-06-13T16:30:00+00:00]>使用工厂模式实现自定义Arrow派生类型，来继承和扩展Arrow的模块API。首先，定义您的类型：
class CustomArrow(arrow.Arrow):
...
... def days_till_xmas(self):
... xmas = arrow.Arrow(self.year, 12, 25)
... if self > xmas:
... xmas = xmas.shift(years=1)
... return (xmas - self).days然后使用工厂：
>>> factory = arrow.ArrowFactory(CustomArrow)
>>> custom = factory.utcnow()
>>> custom
<CustomArrow [2024-06-13T10:02:04.898347+00:00]>
>>> custom.days_till_xmas()
194使用以下标记进行分析和格式化。请注意，它们与strptime的标记不同：
| 标记格式 | 示例 | |
|---|---|---|
| Year | YYYY | 2000, 2001, 2002 … 2012, 2013 |
| YY | 00, 01, 02 … 12, 13 | |
| Month | MMMM | January, February, March …1 |
| MMM | Jan, Feb, Mar …1 | |
| MM | 01, 02, 03 … 11, 12 | |
| M | 1, 2, 3 … 11, 12 | |
| Day of Year | DDDD | 001, 002, 003 … 364, 365 |
| DDD | 1, 2, 3 … 364, 365 | |
| Day of Month | DD | 01, 02, 03 … 30, 31 |
| D | 1, 2, 3 … 30, 31 | |
| Do | 1st, 2nd, 3rd … 30th, 31st | |
| Day of Week | dddd | Monday, Tuesday, Wednesday …2 |
| ddd | Mon, Tue, Wed …2 | |
| d | 1, 2, 3 … 6, 7 | |
| ISO week date | W | 2011-W05-4, 2019-W17 |
| Hour | HH | 00, 01, 02 … 23, 24 |
| H | 0, 1, 2 … 23, 24 | |
| hh | 01, 02, 03 … 11, 12 | |
| h | 1, 2, 3 … 11, 12 | |
| AM / PM | A | AM, PM, am, pm1 |
| a | am, pm1 | |
| Minute | mm | 00, 01, 02 … 58, 59 |
| m | 0, 1, 2 … 58, 59 | |
| Second | ss | 00, 01, 02 … 58, 59 |
| s | 0, 1, 2 … 58, 59 | |
| Sub-second | S… | 0, 02, 003, 000006, 123123123123…3 |
| Timezone | ZZZ | Asia/Baku, Europe/Warsaw, GMT …4 |
| ZZ | -07:00, -06:00 … +06:00, +07:00, +08, Z | |
| Z | -0700, -0600 … +0600, +0700, +08, Z | |
| Seconds Timestamp | X | 1381685817, 1381685817.915482 …5 |
| ms or µs Timestamp | x | 1569980330813, 1569980330813221 |
>>> arw = arrow.utcnow()
>>> arw.format(arrow.FORMAT_ATOM)
'2024-06-13 10:05:40+00:00'
>>> arw.format(arrow.FORMAT_COOKIE)
'Thursday, 13-Jun-2024 10:05:40 UTC'
>>> arw.format(arrow.FORMAT_RSS)
'Thu, 13 Jun 2024 10:05:40 +0000'
>>> arw.format(arrow.FORMAT_RFC822)
'Thu, 13 Jun 24 10:05:40 +0000'
>>> arw.format(arrow.FORMAT_RFC850)
'Thursday, 13-Jun-24 10:05:40 UTC'
>>> arw.format(arrow.FORMAT_RFC1036)
'Thu, 13 Jun 24 10:05:40 +0000'
>>> arw.format(arrow.FORMAT_RFC1123)
'Thu, 13 Jun 2024 10:05:40 +0000'
>>> arw.format(arrow.FORMAT_RFC2822)
'Thu, 13 Jun 2024 10:05:40 +0000'
>>> arw.format(arrow.FORMAT_RFC3339)
'2024-06-13 10:05:40+00:00'
>>> arw.format(arrow.FORMAT_W3C)
'2024-06-13 10:05:40+00:00'解析和格式化时，可以通过将格式字符串中的标记、短语和正则表达式括在方括号中来转义它们。
>>> fmt = "YYYY-MM-DD h [h] m"
>>> arw = arrow.get("2024-06-13 8 h 40", fmt)
>>> arw.format(fmt)
'2024-06-13 8 h 40'
>>> fmt = "YYYY-MM-DD h [hello] m"
>>> arw = arrow.get("2024-06-13 8 hello 40", fmt)
>>> arw
<Arrow [2024-06-13T08:40:00+00:00]>
>>> arw.format(fmt)
'2024-06-13 8 hello 40'
>>> fmt = "YYYY-MM-DD h [hello world] m"
>>> arw = arrow.get("2024-05-16 8 hello world 40", fmt)
>>> arw.format(fmt)
'2024-05-16 8 hello world 40'您还可以通过将正则表达式括在方括号内来转义它们。在下面的示例中，我们使用正则表达式s+来匹配分隔标记的任意数量的空白字符。如果您不提前知道令牌之间的空间数（例如，在日志文件中），这将非常有用。
>>> fmt = r"ddd[\s+]MMM[\s+]DD[\s+]HH:mm:ss[\s+]YYYY"
>>> arrow.get("Thu Jun 16 16:41:45 2024", fmt)
<Arrow [2024-06-16T16:41:45+00:00]>
>>> arrow.get("Thu \tJun 16 16:41:45 2024", fmt)
<Arrow [2024-06-16T16:41:45+00:00]>
>>> arrow.get("Thu Jun 16 16:41:45 2024", fmt)
<Arrow [2024-06-16T16:41:45+00:00]>日期和时间格式的两侧可以用以下列表中的一个标点符号隔开：、.；：！" \ ' [ ] { } ( ) < >`
>>> arrow.get("Cool date: 2024-06-16T09:12:45.123456+04:30.", "YYYY-MM-DDTHH:mm:ss.SZZ")
<Arrow [2024-06-16T09:12:45.123456+04:30]>
>>> arrow.get("Tomorrow (2024-06-16) is Halloween!", "YYYY-MM-DD")
<Arrow [2024-06-16T00:00:00+00:00]>
>>> arrow.get("Halloween is on 2024.06.16.", "YYYY.MM.DD")
<Arrow [2024-06-16T00:00:00+00:00]>
>>> arrow.get("It's Halloween tomorrow (2024-06-16)!", "YYYY-MM-DD")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "C:\Users\LocalCache\local-packages\Python312\site-packages\arrow\api.py", line 91, in get
return _factory.get(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\LocalCache\local-packages\Python312\site-packages\arrow\factory.py", line 292, in get
dt = parser.DateTimeParser(locale).parse(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\LocalCache\local-packages\Python312\site-packages\arrow\parser.py", line 323, in parse
raise ParserMatchError(
arrow.parser.ParserMatchError: Failed to match 'YYYY-MM-DD' when parsing "It's Halloween tomorrow (2024-06-16)!".多余的空白字符（空格、制表符和换行符）可以通过将normalize_whitespace标志传递到arrow.get来自动规范化：
>>> arrow.get('\t \n 2024-06-13T12:30:45.123456 \t \n', normalize_whitespace=True)
<Arrow [2024-06-13T12:30:45.123456+00:00]>
>>> arrow.get('2024-06-13 T \n 12:30:45\t123456', 'YYYY-MM-DD T HH:mm:ss S', normalize_whitespace=True)
<Arrow [2024-06-13T12:30:45.123456+00:00]>