O'Reilly logo

Regular Expressions Cookbook, 2nd Edition by Steven Levithan, Jan Goyvaerts

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

4.7. Validate ISO 8601 Dates and Times

Problem

You want to match dates and/or times in the official ISO 8601 format, which is the basis for many standardized date and time formats. For example, in XML Schema, the built-in date, time, and dateTime types are all based on ISO 8601.

Solution

The ISO 8601 standard defines a wide range of date and time formats. Most applications that use ISO 8601 only use a subset of it. These solutions match the most commonly used ISO 8601 date and time formats. We’ve also added solutions for XML Schema, which is one particular implementation of ISO 8601.

Dates

The following matches a calendar month (e.g., 2008-08). The hyphen is required:

^([0-9]{4})-(1[0-2]|0[1-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby

Named capture makes the regular expression and any code that may reference the capturing groups easier to read:

^(?<year>[0-9]{4})-(?<month>1[0-2]|0[1-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Python uses a different syntax for named capture, adding a P. For brevity, we only show one solution using the Python syntax. All the other solutions using .NET-style named capture can be easily adapted to Python-style named capture in the same way.

^(?P<year>[0-9]{4})-(?P<month>1[0-2]|0[1-9])$
Regex options: None
Regex flavors: PCRE, Python

ISO 8601 allows hyphens to be omitted from calendar dates, making both 2010-08-20 and 20100820 valid representations of the same date. The following regex accounts for this, but also allows for invalid formats like YYYY-MMDD and YYYYMM-DD.

^([0-9]{4})-?(1[0-2]|0[1-9])-?(3[01]|0[1-9]|[12][0-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>[0-9]{4})-?(?<month>1[0-2]|0[1-9])-?↵
(?<day>3[01]|0[1-9]|[12][0-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Calendar date, such as 2008-08-30 or 20080830. The hyphens are optional. This regex uses a capturing group and a backreference to match YYYY-MM-DD or YYYYMMDD, but not YYYY-MMDD or YYYYMM-DD.

^([0-9]{4})(-?)(1[0-2]|0[1-9])\2(3[01]|0[1-9]|[12][0-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>[0-9]{4})(?<hyphen>-?)(?<month>1[0-2]|0[1-9])↵
\k<hyphen>(?<day>3[01]|0[1-9]|[12][0-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Python also uses a different syntax for named backreferences:

^(?P<year>[0-9]{4})(?P<hyphen>-?)(?P<month>1[0-2]|0[1-9])↵
(?P=hyphen)(?<day>3[01]|0[1-9]|[12][0-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Ordinal date (e.g., 2008-243). The hyphen is optional:

^([0-9]{4})-?(36[0-6]|3[0-5][0-9]|[12][0-9]{2}|0[1-9][0-9]|00[1-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>[0-9]{4})-?↵
(?<day>36[0-6]|3[0-5][0-9]|[12][0-9]{2}|0[1-9][0-9]|00[1-9])$
Regex options: None
Regex flavors: .NET, Java 7, PCRE 7, Perl 5.10, Ruby 1.9

Weeks

Week of the year (e.g., 2008-W35). The hyphen is optional:

^([0-9]{4})-?W(5[0-3]|[1-4][0-9]|0[1-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>[0-9]{4})-?W(?<week>5[0-3]|[1-4][0-9]|0[1-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Week date (e.g., 2008-W35-6). The hyphens are optional.

^([0-9]{4})-?W(5[0-3]|[1-4][0-9]|0[1-9])-?([1-7])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>[0-9]{4})-?W(?<week>5[0-3]|[1-4][0-9]|0[1-9])-?(?<day>[1-7])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Times

Hours and minutes (e.g., 17:21). The colon is optional:

^(2[0-3]|[01][0-9]):?([0-5][0-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<hour>2[0-3]|[01][0-9]):?(?<minute>[0-5][0-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Hours, minutes, and seconds (e.g., 17:21:59). The colons are optional:

^(2[0-3]|[01][0-9]):?([0-5][0-9]):?([0-5][0-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<hour>2[0-3]|[01][0-9]):?(?<minute>[0-5][0-9]):?↵
(?<second>[0-5][0-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Time zone designator (e.g., Z, +07 or +07:00). The colons and the minutes are optional:

^(Z|[+-](?:2[0-3]|[01][0-9])(?::?(?:[0-5][0-9]))?)$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby

Hours, minutes, and seconds with time zone designator (e.g., 17:21:59+07:00). All the colons are optional. The minutes in the time zone designator are also optional:

^(2[0-3]|[01][0-9]):?([0-5][0-9]):?([0-5][0-9])↵
(Z|[+-](?:2[0-3]|[01][0-9])(?::?(?:[0-5][0-9]))?)$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<hour>2[0-3]|[01][0-9]):?(?<minute>[0-5][0-9]):?(?<second>[0-5][0-9])↵
(?<timezone>Z|[+-](?:2[0-3]|[01][0-9])(?::?(?:[0-5][0-9]))?)$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Date and time

Calendar date with hours, minutes, and seconds (e.g., 2008-08-30 17:21:59 or 20080830 172159). A space is required between the date and the time. The hyphens and colons are optional. This regex matches dates and times that specify some hyphens or colons but omit others. This does not follow ISO 8601.

^([0-9]{4})-?(1[0-2]|0[1-9])-?(3[01]|0[1-9]|[12][0-9])↵
(2[0-3]|[01][0-9]):?([0-5][0-9]):?([0-5][0-9])$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>[0-9]{4})-?(?<month>1[0-2]|0[1-9])-?↵
(?<day>3[01]|0[1-9]|[12][0-9])(?<hour>2[0-3]|[01][0-9])↵
:?(?<minute>[0-5][0-9]):?(?<second>[0-5][0-9])$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

A more complicated solution is needed if we want to match date and time values that specify either all of the hyphens and colons, or none of them. The cleanest solution is to use conditionals. But only some flavors support conditionals.

^([0-9]{4})(-)?(1[0-2]|0[1-9])(?(2)-)(3[01]|0[1-9]|[12][0-9])↵
(2[0-3]|[01][0-9])(?(2):)([0-5][0-9])(?(2):)([0-5][0-9])$
Regex options: None
Regex flavors: .NET, PCRE, Perl, Python
^(?<year>[0-9]{4})(?<hyphen>-)?(?<month>1[0-2]|0[1-9])↵
(?(hyphen)-)(?<day>3[01]|0[1-9]|[12][0-9])(?<hour>2[0-3]|[01][0-9])↵
(?(hyphen):)(?<minute>[0-5][0-9])(?(hyphen):)(?<second>[0-5][0-9])$
Regex options: None
Regex flavors: .NET, PCRE 7, Perl 5.10
^(?P<year>[0-9]{4})(?P<hyphen>-)?(?P<month>1[0-2]|0[1-9])↵
(?(hyphen)-)(?P<day>3[01]|0[1-9]|[12][0-9])(?P<hour>2[0-3]|[01][0-9])↵
(?(hyphen):)(?P<minute>[0-5][0-9])(?(hyphen):)(?P<second>[0-5][0-9])$
Regex options: None
Regex flavors: PCRE, Perl 5.10, Python

If conditionals are not available, then we have to use alternation to spell out the alternatives with and without delimiters.

^(?:([0-9]{4})-?(1[0-2]|0[1-9])-?(3[01]|0[1-9]|[12][0-9])↵
(2[0-3]|[01][0-9]):?([0-5][0-9]):?([0-5][0-9])|↵
([0-9]{4})(1[0-2]|0[1-9])(3[01]|0[1-9]|[12][0-9])↵
(2[0-3]|[01][0-9])([0-5][0-9])([0-5][0-9]))$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby

XML Schema dates and times

The date and time types defined in the XML Schema standard are based on the ISO 8601 standard. The date types allow negative years for years before the start of the calendar (B.C. years). It also allows for years with more than four digits, but not for years with fewer than four digits. Years with more than four digits must not have leading zeros. If you only want to allow years with four digits as in the preceding solutions, remove -?(?:[1-9][0-9]*)? from the following solutions.

Date, with optional time zone (e.g., 2008-08-30 or 2008-08-30+07:00). Hyphens are required. This is the XML Schema date type:

^(-?(?:[1-9][0-9]*)?[0-9]{4})-(1[0-2]|0[1-9])-(3[01]|0[1-9]|[12][0-9])↵
(Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>-?(?:[1-9][0-9]*)?[0-9]{4})-(?<month>1[0-2]|0[1-9])-↵
(?<day>3[01]|0[1-9]|[12][0-9])↵
(?<timezone>Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Time, with optional fractional seconds and time zone (e.g., 01:45:36 or 01:45:36.123+07:00). There is no limit on the number of digits for the fractional seconds. This is the XML Schema time type:

^(2[0-3]|[01][0-9]):([0-5][0-9]):([0-5][0-9])(\.[0-9]+)?↵
(Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<hour>2[0-3]|[01][0-9]):(?<minute>[0-5][0-9]):(?<second>[0-5][0-9])↵
(?<frac>\.[0-9]+)?(?<timezone>Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Date and time, with optional fractional seconds and time zone (e.g., 2008-08-30T01:45:36 or 2008-08-30T01:45:36.123Z). This is the XML Schema dateTime type:

^(-?(?:[1-9][0-9]*)?[0-9]{4})-(1[0-2]|0[1-9])-(3[01]|0[1-9]|[12][0-9])↵
T(2[0-3]|[01][0-9]):([0-5][0-9]):([0-5][0-9])(\.[0-9]+)?↵
(Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$
Regex options: None
Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby
^(?<year>-?(?:[1-9][0-9]*)?[0-9]{4})-(?<month>1[0-2]|0[1-9])-↵
(?<day>3[01]|0[1-9]|[12][0-9])T(?<hour>2[0-3]|[01][0-9]):↵
(?<minute>[0-5][0-9]):(?<second>[0-5][0-9])(?<ms>\.[0-9]+)?↵
(?<timezone>Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$
Regex options: None
Regex flavors: .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, Ruby 1.9

Discussion

ISO 8601 defines a wide range of date and time formats. The regular expressions presented here cover the most common formats, but most systems that use ISO 8601 only use a subset. For example, in XML Schema dates and times, the hyphens and colons are mandatory. To make hyphens and colons mandatory, simply remove the question marks after them. To disallow hyphens and colons, remove the hyphens and colons along with the question mark that follows them. Do watch out for the noncapturing groups, which use the (?:) syntax. If a question mark and a colon follow an opening parenthesis, those three characters open a noncapturing group.

We put parentheses around all the number parts of the regexes. That makes it easy to retrieve the numbers for the years, months, days, hours, minutes, seconds, and time zones. Recipe 2.9 explains how parentheses create capturing groups. Recipe 3.9 explains how you can retrieve the text matched by those capturing groups in procedural code.

For most regexes, we also show an alternative using named capture. Some of these date and time formats may be unfamiliar to you or your fellow developers. Named capture makes the regex easier to understand. .NET, Java 7, XRegExp, PCRE 7, Perl 5.10, and Ruby 1.9 support the (?<name>) syntax used in the solutions in this recipe. All versions of PCRE and Python covered in this book support the alternative (?P<name>) syntax, which adds a P. See Recipes 2.11 and 3.9 for details.

The number ranges in all the regexes are strict. For example, the calendar day is restricted between 01 and 31. You’ll never end up with day 32 or month 13. None of the regexes here attempts to exclude invalid day and month combinations, such as February 31st; Recipe 4.5 explains how you can deal with that.

The regular expressions, except those in the XML Schema subsection, make the individual hyphens and colons optional. This does not follow ISO 8601 exactly. For example, 1733:26 is not a valid ISO 8601 time, but will be accepted by the time regexes. Requiring all hyphens and colons to be present or omitted at the same time makes your regex quite a bit more complex.

If the delimiters are all the same, we can do this quite easily using a capturing group for the first delimiter and backreferences for the remaining delimiters. The “dates” subsection of the “Solution” section shows an example. For the first hyphen, we use (-?), (?<hyphen>-?) or (?P<hyphen>-?) to match an optional hyphen and capture it into a named or numbered group. If the hyphen was omitted, the capturing group stores the zero-length string. The question mark that makes the hyphen optional must be inside the group. If we made the group itself optional, then backreferences to that group would always fail to match if the hyphen was not matched, as the group would not have participated in the match at all. For the remaining hyphens, we use \2, \k<hyphen>, or (?P=hyphen) to match the same text that was matched by the capturing group, which is either a hyphen or nothing at all, depending on whether the first hyphen was matched or not. When using numbered capture, make sure to use the correct number for the backreference.

If the delimiters are different, such as when matching a single string with both a date and a time, the solution is more complex. The “date and time” subsection shows an example. This time, we use (-)?, (?<hyphen>-)? or (?P<hyphen>-)? to match the hyphen. Now the question mark is outside the capturing group so that it will not participate in the match at all when the hyphen is omitted. This allows us to use the capturing group with a conditional. (?(2)-) matches a hyphen and (?(2):) matches a colon if the second capturing group participated in the match. The conditionals have no alternative, which means they will match nothing at all (but still succeed) when the second capturing group did not participate in the match. (?(hyphen)-) and (?(hyphen):) do the same using named capture.

Only some flavors support conditionals. If conditionals are not available, the only solution is to use alternation to spell out the two alternatives with and without delimiters. The disadvantage of this solution is that it results in two capturing groups for each part of the date and time. Only one of the two sets of capturing groups will participate in the match. Code that uses this regex will have to check both groups.

See Also

This chapter has several other recipes for matching dates and times. Recipes 4.4 and 4.5 show how to validate traditional date formats. Recipe 4.6 shows how to validate traditional time formats.

Techniques used in the regular expressions in this recipe are discussed in Chapter 2. Recipe 2.3 explains character classes. Recipe 2.5 explains anchors. Recipe 2.8 explains alternation. Recipe 2.9 explains grouping. Recipe 2.10 explains backreferences. Recipe 2.11 explains named capturing groups. Recipe 2.12 explains repetition. Recipe 2.17 explains conditionals.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required