Re: Pattern Matching question - PG 9.6

2017-05-15 16:10 GMT+12:00 David G. Johnston <david(dot)g(dot)johnston(at)gmail(dot)com>:

> On Sunday, May 14, 2017, Patrick B <patrickbakerbr(at)gmail(dot)com> wrote:
>
>>
>> Demo: http://dbfiddle.uk/?rdbms=postgres_9.6&fiddle=3c3a3f870eb4d0
>> 02c5b4200042b25669
>> <http://dbfiddle.uk/?rdbms=postgres_9.6&fiddle=c2fbb7da5a2397f7cda5126ed239c080>
>>
>> The rows that I should be getting are:
>>
>> 5 /testfile/client/10/attachment/1000/master/ 10
>>
>> 7 /testfile/client/10/attachment/unassigned/file/1001/master 10
>>
>> 8 /testfile/client/10/attachment/unassigned/file/1002/master 10
>>
>> What am I doing wrong?
>>
>
> Without you explaining why 6 and 9 are invalid it's impossible to say how
> you should modify your regex to exclude them. You may find positive and
> negative look-ahead useful though.
>
> David J.
>

I thought I have already explained it. Here it goes again. Demo page is:
http://dbfiddle.uk/?rdbms=postgres_9.6&fiddle=ea61e7e1859bdb7f297f853a9dc0e3
d0

As you can see, there is a lot of variations for the same file_id (1000).
File_id (1001/1002) is a new unassigned file, different from the others.

I wanna be able to get ONLY the 'master' variation (
/testfile/client/10/attachment/1000/master/ ) and the unassigned files
variations [if any]
(/testfile/client/10/attachment/unassigned/file/1001/master |
/testfile/client/10/attachment/unassigned/file/1002/master).

So on the demo above, only id IN (5,9,10) are valid for me. The SELECT that
I used as an example is not returning me ONLY the data I need, instead, it
is returning (almost) everything.

To summarize: I wanna use a pattern matching the only returns these rows:

/testfile/client/10/attachment/1000/master/
/testfile/client/10/attachment/unassigned/file/1001/master
/testfile/client/10/attachment/unassigned/file/1002/master

What can I do to fix it?
Thanks
P.