How do I split a date that is of the form YYYYMMDD into its constituents?
my ($yyyy, $mm, $dd) = $date =~ /(\4d+)(\2d+)(\2d+)/;
views:
153answers:
5
+5
A:
my ($year, $month, $day) = $date =~ /^(\d{4})(\d{2})(\d{2})\z/
or die "bad date: $date";
ysth
2010-07-02 22:30:33
What is the z for?
syker
2010-07-02 22:31:42
maybe `$` was meant instead of `\z`?
jdehaan
2010-07-02 22:32:58
What would $ mean then?
syker
2010-07-02 22:34:48
Could i use o instead? I'm using it two other times (not in a loop though)
syker
2010-07-02 22:38:37
@syker: please see [perldoc perlre](http://perldoc.perl.org/perlre.html) for details about regular expression syntax (e.g. what the z, $, \d and other components mean).
Ether
2010-07-02 22:40:39
@syker: @jdehaan: `$` would mean either the end of the string or \n and then the end of the string. Since the specified format shouldn't have a \n, I used \z
ysth
2010-07-03 00:10:46
@syker: o isn't relevant here - I don't understand what you are proposing it replace
ysth
2010-07-03 00:11:58
+4
A:
#!/usr/bin/perl -w
use strict;
sub main{
my $date = "some text with the numbers 2010063011 and more text";
print "Input Date: $date\n";
my ($year, $month, $day) = $date =~ /\b(\d{4})(\d{2})(\d{2})\b/;
print qq{
Date: $date
Year: $year
Month: $month
Day: $day\n} if (defined $year && defined $month && defined $day);
}
main();
notice this will look for the first date in the regex, it won't work with 2010063011 because it's not a date, but it will work with 20100630, which is what you want.
vol7ron
2010-07-02 22:55:00
your method vs the unpack method makes sense, thanks for adding that comment in the `unpack` answer
Armando
2010-07-04 17:33:29
+11
A:
my ($year, $month, $day) = unpack "A4A2A2", $date;
pack and unpack are underused builtins that can be used to great power.
Leon Timmermans
2010-07-02 22:55:09
unpack is the fastest perl has to offer, especially for fixed-width data.
vol7ron
2010-07-02 23:05:59
can you do error checking with unpack? for ex: have an or 'die "Wrong format"' statement after that?
syker
2010-07-02 23:11:51
@ysth: using a regexp doesn't properly validate either. It will allow 9999-99-99 as a date, even though that clearly isn't a valid date.
Leon Timmermans
2010-07-03 10:34:43
It depends on how `$date` is being stored, `unpack` is only good for fixed width. Given my example, if there's any other text other than 8 digits (even spaces), then `unpack` will break. That's why my answer adds some pre-validation, using `\b`.
vol7ron
2010-07-03 15:18:51
@Leon Timmermans: yes, but it's a good first step; at least you know your variables are numeric can treat them as such.
ysth
2010-07-04 20:42:25
A:
Note that \d matches any Unicode digits, not just latin decimal digits.
So if you want to do input validation use '[0-9]' instead of '\d'.
dolmen
2010-07-05 09:27:38
A:
Whenever I need to work with dates I use the DateTime module. You can grab it from CPAN.
Proclus
2010-07-05 13:22:43