Horton

Thread

  1. How best to work around the issue - regex string cannot contain brackets

    Shaozhong SHI <shishaozhong@gmail.com> — 2022-02-03T16:53:18Z

    One would consider the following would work, but it did not because the
brackets.
select regexp_matches('Department for Transport (Parking)', 'Department for
Transport (Parking)', 'g')

Can anyone enlighten me?

Regards,

David

  2. Re: How best to work around the issue - regex string cannot contain brackets

    Christophe Pettus <xof@thebuild.com> — 2022-02-03T16:58:45Z

    
> On Feb 3, 2022, at 08:53, Shaozhong SHI <shishaozhong@gmail.com> wrote:
> 
> One would consider the following would work, but it did not because the brackets.
> select regexp_matches('Department for Transport (Parking)', 'Department for Transport (Parking)', 'g')
> 
> Can anyone enlighten me?

You escape the ()s with a backslash:

xof=# select regexp_matches('Department for Transport (Parking)', 'Department for Transport \(Parking\)', 'g');
             regexp_matches             
----------------------------------------
 {"Department for Transport (Parking)"}
(1 row)

  3. Re: How best to work around the issue - regex string cannot contain brackets

    David G. Johnston <david.g.johnston@gmail.com> — 2022-02-03T17:04:42Z

    On Thu, Feb 3, 2022 at 9:58 AM Christophe Pettus <xof@thebuild.com> wrote:

>
>
> > On Feb 3, 2022, at 08:53, Shaozhong SHI <shishaozhong@gmail.com> wrote:
> >
> > One would consider the following would work, but it did not because the
> brackets.
> > select regexp_matches('Department for Transport (Parking)', 'Department
> for Transport (Parking)', 'g')
> >
> > Can anyone enlighten me?
>

Have you tried reading a book or some tutorials on RegExes?  I'll admit our
documentation is probably not the best resource out there to actually learn
the language.


> You escape the ()s with a backslash:
>

More generally this behavior this documented as "\k"

https://www.postgresql.org/docs/current/functions-matching.html#FUNCTIONS-POSIX-REGEXP

David J.

  4. Re: How best to work around the issue - regex string cannot contain brackets

    hubert depesz lubaczewski <depesz@depesz.com> — 2022-02-04T12:55:58Z

    On Thu, Feb 03, 2022 at 04:53:18PM +0000, Shaozhong SHI wrote:
> One would consider the following would work, but it did not because the
> brackets.
> select regexp_matches('Department for Transport (Parking)', 'Department for
> Transport (Parking)', 'g')
> Can anyone enlighten me?

Perhaps you don't want regexp matching, but simple equality or
substring match?

depesz

  5. Re: How best to work around the issue - regex string cannot contain brackets

    Shaozhong SHI <shishaozhong@gmail.com> — 2022-02-04T14:00:55Z

    It appears that the following regex work differently.

Why \d and [\d] are different?

[A-PR-UWYZ]\d{1,2} and [A-PR-UWYZ][\d]{1,2}

Regards,

David

On Thu, 3 Feb 2022 at 17:04, David G. Johnston <david.g.johnston@gmail.com>
wrote:

> On Thu, Feb 3, 2022 at 9:58 AM Christophe Pettus <xof@thebuild.com> wrote:
>
>>
>>
>> > On Feb 3, 2022, at 08:53, Shaozhong SHI <shishaozhong@gmail.com> wrote:
>> >
>> > One would consider the following would work, but it did not because the
>> brackets.
>> > select regexp_matches('Department for Transport (Parking)', 'Department
>> for Transport (Parking)', 'g')
>> >
>> > Can anyone enlighten me?
>>
>
> Have you tried reading a book or some tutorials on RegExes?  I'll admit
> our documentation is probably not the best resource out there to actually
> learn the language.
>
>
>> You escape the ()s with a backslash:
>>
>
> More generally this behavior this documented as "\k"
>
>
> https://www.postgresql.org/docs/current/functions-matching.html#FUNCTIONS-POSIX-REGEXP
>
> David J.
>
>

  6. Re: How best to work around the issue - regex string cannot contain brackets

    Steve Midgley <science@misuse.org> — 2022-02-04T17:14:39Z

    On Fri, Feb 4, 2022 at 6:01 AM Shaozhong SHI <shishaozhong@gmail.com> wrote:

> It appears that the following regex work differently.
>
> Why \d and [\d] are different?
>
> [A-PR-UWYZ]\d{1,2} and [A-PR-UWYZ][\d]{1,2}
>
>>
>>
This is getting into regex stuff, where maybe stackoverflow is a better
resource? But when you put characters into brackets, you are telling regex
to search for each character represented in the bracket. So [\d] is looking
for any single character that is either a \ or a d character. Outside of
brackets, regex evaluates \d as any digit. For US English charset [0-9] is
equivalent to \d I believe.

  7. Re: How best to work around the issue - regex string cannot contain brackets

    David G. Johnston <david.g.johnston@gmail.com> — 2022-02-04T17:24:24Z

    On Fri, Feb 4, 2022 at 10:14 AM Steve Midgley <science@misuse.org> wrote:

>
>
> On Fri, Feb 4, 2022 at 6:01 AM Shaozhong SHI <shishaozhong@gmail.com>
> wrote:
>
>> It appears that the following regex work differently.
>>
>> Why \d and [\d] are different?
>>
>> [A-PR-UWYZ]\d{1,2} and [A-PR-UWYZ][\d]{1,2}
>>
>
Show your work!


> But when you put characters into brackets, you are telling regex to search
> for each character represented in the bracket. So [\d] is looking for any
> single character that is either a \ or a d character. Outside of brackets,
> regex evaluates \d as any digit.
>

This is simply wrong.  Test it.

David J.

  8. Re: How best to work around the issue - regex string cannot contain brackets

    Steve Midgley <science@misuse.org> — 2022-02-04T17:29:53Z

    On Fri, Feb 4, 2022 at 9:24 AM David G. Johnston <david.g.johnston@gmail.com>
wrote:

> On Fri, Feb 4, 2022 at 10:14 AM Steve Midgley <science@misuse.org> wrote:
>
>>
>>
>> On Fri, Feb 4, 2022 at 6:01 AM Shaozhong SHI <shishaozhong@gmail.com>
>> wrote:
>>
>>> It appears that the following regex work differently.
>>>
>>> Why \d and [\d] are different?
>>>
>>> [A-PR-UWYZ]\d{1,2} and [A-PR-UWYZ][\d]{1,2}
>>>
>>
> Show your work!
>
>
>> But when you put characters into brackets, you are telling regex to
>> search for each character represented in the bracket. So [\d] is looking
>> for any single character that is either a \ or a d character. Outside of
>> brackets, regex evaluates \d as any digit.
>>
>
>
Apologies - I was relying on stackoverflow and my prior experiences with
regex engines that don't honor shorthand characters inside brackets. I
should have tested postgres latest before responding. Thanks for the
correction. (
https://stackoverflow.com/questions/46020936/perl-like-shorthand-character-class-not-working-inside-bracket-expression
)