How to use REGEXP in MYSQL to match dates IN array? - php

Question

I am trying to match the year from the publication_date of a table in a lot of years. I am using laravel (PHP).
I have written something like:
if($this->request->has('publication_years')){
$years = implode('-[0-9]{2}-[0-9]{2}|', explode(',', $this->request->get('publication_years'))) . '-[0-9]{2}-[0-9]{2}';
$model = $model->whereExists(function ($query) use ($years)
{
$query->select(DB::raw(1))
->from('patent')
->whereRaw('patent.id = document.documentable_id')
->whereRaw('document.documentable_type = "patent"')
->whereRaw('(patent.publication_date REGEXP (' . $years . '))');
});
}
The years come like ?publication_years=2015,2016. I explode them use the , delimiter and then implode them to make the REGEXP pattern in a way it matches the date. The dump of the years after this is like 2015-[0-9]{2}-[0-9]{2}|2016-[0-9]{2}-[0-9]{2}.
This works at http://regexr.com/ but I am still getting error so I am guessing there some other syntax for MYSQL. The error that I see is:
SQLSTATE[42000]: Syntax error or access violation: 1064 You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near '[0-9]{2}-[0-9]{2}|2016-[0-9]{2}-[0-9]{2})))' at line 1 (SQL: select count(*) as aggregate from `document` where exists (select * from `folder` where `document`.`folder_id` = `folder`.`id` and (exists (select * from `user` inner join `accessible_by_user` on `user`.`id` = `accessible_by_user`.`user_id` where `accessible_by_user`.`accessible_id` = `folder`.`id` and `accessible_by_user`.`accessible_type` = folder and `user_id` = 1) or exists (select * from `role` inner join `accessible_by_role` on `role`.`id` = `accessible_by_role`.`role_id` where `accessible_by_role`.`accessible_id` = `folder`.`id` and `accessible_by_role`.`accessible_type` = folder and `role_id` in (1)))) and exists (select 1 from `patent` where patent.id = document.documentable_id and document.documentable_type = "patent" and (patent.publication_date REGEXP (2015-[0-9]{2}-[0-9]{2}|2016-[0-9]{2}-[0-9]{2}))))
What am I doing wrong? How can I correct this?

Answer 1

Here is a way to prep the regex for using in MySQL:
$this_request_get_publication_years = "2015,2016,2017";
$years = '^(' . implode('|', explode(',',
$this_request_get_publication_years)) . ')-[0-9]{2}-[0-9]{2}';
echo "(patent.publication_date REGEXP '" . $years . "')";
// => (patent.publication_date REGEXP '^(2015|2016|2017)-[0-9]{2}-[0-9]{2}')
The REGEXP '^(2015|2016|2017)-[0-9]{2}-[0-9]{2}' will only fetch the records that:
^ - at the start of the string, have...
(2015|2016|2017) - either of the 3 years (and more are supported, surely)
-[0-9]{2} - a hyphen followed with 2 digits
-[0-9]{2} - ibid.
The latter two subpatterns can be shrunk into (-[0-9]{2}){2}.