Re: back references using regex

That doesn't seem to work with digits

SELECT substring('12.00.00.34' FROM $$((\d+)\.\2)$$);
 or
SELECT substring('12.00.00.34' FROM $$((\w+)\.\2)$$);

but works with strings

SELECT substring('abc.foo.foo.xyz' FROM
$$((\w+)\.\2)$$);

What I basically want to do is have a slice function
like Python, where I can slice out items from a \s, \.
or \n\n separated list.

Where I could start it at a certain point and end it
at another. Like slicing out paragraph 3-6 (delimiter
\n\n) or the 2nd-6th sentence in a article (delimiter
\.). That is what I am trying to do. I know if I can
figure how to get that working I can figure out how to
deal with extending it/handling misc. characters/etc
in pgsql. I only need a working vanilla example.
Thanks for all you help.


--- Michael Fuhr <mike@fuhr.org> wrote:
> SELECT substring('abc.foo.foo.xyz' FROM
> '(([[:alpha:]]+)\\.\\2)');
>  substring
> -----------
>  foo.foo
> (1 row)
>
> That is, one or more alphabetic characters followed
> by a dot followed
> by the same set of characters (this is a simplistic
> example: it would
> also match 'foo.oog' and return 'oo.oo').
>
> Note that the back reference is \2 because it refers
> to the inner
> set of parentheses (i.e., the subexpression with the
> second opening
> parenthesis); the outer set is used here for
> capturing.  And again,
> note the escaped backslashes because we're using
> ordinary quotes.
> With dollar quotes the above query would be:
>
> SELECT substring('abc.foo.foo.xyz' FROM
> $$(([[:alpha:]]+)\.\2)$$);
>
> --
> Michael Fuhr
>


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com