Re: Need regexp help (seemingly simple problem)
Date: Fri, 23 Oct 2009 12:12:37 -0700
Message-ID: <26fdee6e0910231212w57a5343fy7f728f00aaf72145_at_mail.gmail.com>
Thanks Rich,
Yes, I was looking for the equiv of Perl look-around functionality.Sorry to hear it's not there. Thanks for the reference. Mike
On Fri, Oct 23, 2009 at 11:53 AM, Rich Jesse < rjoralist_at_society.servebeer.com> wrote:
> Hey Mike,
>
> > First of all I would like to say that finding detailed information about
> > Oracle's implementation of regular expressions is next to impossible.
> Either
> > that, or I am looking in all the wrong places. So, links to any complete
> > online documentation would be appreciated. I found plenty of 'basic
> > explanation' but there has got to be more.
>
> This site might help:
>
> http://www.regular-expressions.info/oracle.html
>
> ...although it's lacking in version-specific information about what
> Oracle's
> implemented. There's also a link at the bottom of the page for a book from
> Amazon.
>
> > I could easily use INSTR and SUBSTR to do this, but my goal is to solve
> the
> > problem using a single REGEXP_SUBSTR statement. Also, I have used the
> > pattern 'a[[:digit:]]+' which results in 'a12345'. This is almost right,
> but
> > I don't want the 'a' to be returned.
>
> I don't think that's possible with Oracle's regex. It's possible in Perl,
> but only in 5.10 and up from what I've found. This is as close as I could
> get:
>
> select regexp_substr('2B4a12345Z2lr7','[[:digit:]]+',
> regexp_instr('2B4a12345Z2lr7','a'))
> from dual;
>
> Yes, the REGEXP_INSTR is overkill as INSTR would work just as well and is
> most likely faster for this particular use.
>
> Here's the doc that describes a possible "true" answer in regex, although
> it
> does not appear to be possible in Oracle's regex.
>
> HTH!
>
> Rich
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
-- http://www.freelists.org/webpage/oracle-lReceived on Fri Oct 23 2009 - 14:12:37 CDT