In an Oracle Database 10g Enterprise Edition Release 10.1.0.2.0 server,
REGEXP_SUBSTR() always returns NULL. E.g.:

SELECT REGEXP_SUBSTR('1234,5678', '\d+', 1, 1)
FROM DUAL

In my local 10g XE, however, it returns 1234 as expected.

Is it me or is it the server?


--
-- http://alvaro.es - Álvaro G. Vicario - Burgos, Spain
-- Mi sitio sobre programación web: http://borrame.com
-- Mi web de humor satinado: http://www.demogracia.com
--

On Jun 25, 6:29*am, "Álvaro G. Vicario"
<alvaro.NOSPAMTH... (AT) demogracia (DOT) com> wrote:
It may be the server if you are, indeed, comparing XE (10.2.0.1) to a
10.1.0.2 release. You're certain you've posted the version for the
Enterprise Edition correctly?


David Fitzjarrell

ddf escribió:
Absolutely: SELECT BANNER FROM v$version + clipboard.

If it was a legacy edition that did not support REGEXP_SUBSTR() I'd
expect an error message.

Is it possible to cancel, disabled or override REGEXP_SUBSTR?


--
-- http://alvaro.es - Álvaro G. Vicario - Burgos, Spain
-- Mi sitio sobre programación web: http://borrame.com
-- Mi web de humor satinado: http://www.demogracia.com
--

On Jun 25, 7:52*am, "Álvaro G. Vicario"
<alvaro.NOSPAMTH... (AT) demogracia (DOT) com> wrote:
I can duplicate the expected result on 10.2.0.3 running on AIX 5.3L.
Does anyone have a 10.1 version that can be used to duplicate the OP
issue?

I have posted to another forum to see if anyone with a 10.1 system can
duplicate the issue. I suspect a patch level specific issue, but
there could be another cause.

HTH -- Mark D Powell --

On Jun 25, 6:52*am, "Álvaro G. Vicario"
<alvaro.NOSPAMTH... (AT) demogracia (DOT) com> wrote:
Things have changed between 10.1 and 10.2:

10;1 regular expression support --
http://download.oracle.com/docs/cd/B...01.htm#i690819

10.2 regular expression support --
http://download.oracle.com/docs/cd/B...01.htm#i690819

These differences may result in the differences you're seeing.


David Fitzjarrell

ddf schrieb:
As David already suggested, perl influenced extensions came in 10.2, the
list is at
http://download.oracle.com/docs/cd/B...p_posix003.htm

On 10.1 one should use instead the posix syntax
REGEXP_SUBSTR('1234,5678', '[[:digit:]]+', 1, 1) or
REGEXP_SUBSTR('1234,5678', '[0-9]+', 1, 1) or

Best regards

Maxim

--
Why make things difficult, when it is possible to make them cryptic
and totally illogical, with just a little bit more effort?

Aksel Peter Jørgensen

Maxim Demenko escribió:
Oh my... That was difficult to spot! I stared at both documents trying
to figure out what had changed, with no luck xD

Of course, that was it: \d was looking for a literal '\d'.

Something I miss in Oracle reference is that it doesn't explain when a
function became available or suffered changes; at least, not in the
function definition. Oh well, thank you everybody. [0-9] is doing the
job just as fine as \d.


--
-- http://alvaro.es - Álvaro G. Vicario - Burgos, Spain
-- Mi sitio sobre programación web: http://borrame.com
-- Mi web de humor satinado: http://www.demogracia.com
--