urlSegment regex year-year

[cb2004]

I must admit, regex is certainly not my strong point. What would I have to put in to allow for an urlSegment of say 2015-2016?

Thanks people.

[cb2004]

I rock at regex

regex:^[0-9-]+$

[netcarver]

Hi @cb2004

That's a fine start However, it will also match "-", "--", "---", "-2000", "-0", "-1000-1", "1---", "31415926" etc. If you are expecting a 4 digit start and end year separated with a single dash then you will need to use a more restrictive/specific regular expression.

Assuming the start year is in the 1900 to 2099 range, and the end year in the 1900 to 2999 range, you'd need something more like this;

Quote

^(19|20)[0-9]{2}-(19|2[0-9])[0-9]{2}$

Also, if you are going to be splitting this string up into the start and end year, you might want to use a urlSegment for the start and a separate one for the end year. If that is the case, you'd use ^(19|20)[0-9]{2}$ for the start urlSegment and ^(19|2[0-9])[0-9]{2}$ for the end urlSegment.

Hope that helps!

Edited September 29, 2016 by netcarver
I even got my regex wrong!

[adrian]

Here's a nice article with some examples if you need to allow dates before 1900

http://www.x-note.co.uk/javascript-regex-validate-years-in-range/

Also don't forget about some of the great regex testing tools:

https://leaverou.github.io/regexplained/
https://regex101.com/
http://www.phpliveregex.com/

 

[netcarver]

I agree with Adrian: the regex101 site is very useful!

[Sergio]

Why not a simpler one, like:

^\d{4}-\d{4}$

?

Does anybody see a problem with this approach?

[adrian]

1 minute ago, Sérgio said:

Why not a simpler one, like:

^\d{4}-\d{4}$

?

Does anybody see a problem with this approach?

Probably nothing wrong - I guess it depends on how determined you are to make sure they are valid years. Yes 9155 is a valid year, but it if was an accidental transposition of the 9 and 1, then it might be nice to exclude it if you are only expecting years plus or minus a hundred of today (or whatever is appropriate).

[Sergio]

5 minutes ago, adrian said:

Probably nothing wrong - I guess it depends on how determined you are to make sure they are valid years. Yes 9155 is a valid year, but it if was an accidental transposition of the 9 and 1, then it might be nice to exclude it if you are only expecting years plus or minus a hundred of today (or whatever is appropriate).

 

 

Agree, but wouldn't be better to check this after the regex has matched, something like this, for instance:

$year_url_segment = $sanitizer->selectorValue($input->urlSegment);
$years_on_database = range(date("Y"),2007); //Range of years from 2007 to present
if (in_array($year_url_segment, $years_on_database) ) {
	//find pages in that year
}

Off course I'm considering the url with just one year, not a period.

Edited September 30, 2016 by Sérgio

[adrian]

3 minutes ago, Sérgio said:

Agree, but wouldn't be better to check this after the regex has matched, something like this, for instance:

$year_url_segment = $sanitizer->selectorValue($input->urlSegment);
$years_on_database = range(date("Y"),2007); //Range of years from 2007 to present
if (in_array($year_url_segment, $years_on_database) ) {
	//find pages in that year
}

Off course I'm considering the url with just one year, not a period.

Yeah, maybe that would be more efficient, but wouldn't you also want to throw a 404 if it didn't match match your required range? Using a more specific regex in the "Which URL segments do you want to allow?" would take care of this automatically. I am no expert in the efficiencies of regex matching, but do you think it's worth having a simpler regex to save the additional checks in the code? I guess you'd need to test the two and see

[netcarver]

@Sérgio There's nothing wrong with that approach, \d{4} should work fine. I guess it also comes down to where you want to codify the rules - in the regex, or in the PHP, or some combination of both.

[LostKobrakai]

I think i'd probably go with this in the regex (if it's not for historical data)

(1|2)\d{3}-(1|2)\d{3}

For something like a blog there's hardly need for years before 1000 and beyond 2999