Regular Expressions; return a buried match
in [Oracle]
|
Prev: New dbTrends Software Product for Oracle AWR and STATSPACK Released
Next: How to use REGEXP_SUBSTR to extract owner/obj/part from full obj name ?
From: Walt on 4 Feb 2010 15:27 cate wrote: > select REGEXP_SUBSTR(encodedField, '^\d{4}(\d\d)') from table; > > I want the 5th and 6th digit. Can Oracle do this? If the 5th and 6th digit have meaning on their own, then they should have been modeled as their own field(s) instead of concatenated into a composite field. That said, yes, one can extract them with simple expressions. Several examples have already been given. I don't see why regular expressions are necessary, but they will work. SUBSTR should work, unless the data is more hinky than you've let on. //Walt
From: cate on 4 Feb 2010 19:17 On Feb 4, 2:27 pm, Walt <walt_ask...(a)SHOESyahoo.com> wrote: > cate wrote: > > select REGEXP_SUBSTR(encodedField, '^\d{4}(\d\d)') from table; > > > I want the 5th and 6th digit. Can Oracle do this? > > If the 5th and 6th digit have meaning on their own, then they should > have been modeled as their own field(s) instead of concatenated into a > composite field. > > That said, yes, one can extract them with simple expressions. Several > examples have already been given. I don't see why regular expressions > are necessary, but they will work. SUBSTR should work, unless the data > is more hinky than you've let on. > > //Walt Thanks all. I think we'll surrender and let perl handle this. Walt, I'll forward your comment to the client. :-)
From: Álvaro G. Vicario on 5 Feb 2010 06:28
El 04/02/2010 21:27, Walt escribi�/wrote: > cate wrote: >> select REGEXP_SUBSTR(encodedField, '^\d{4}(\d\d)') from table; >> >> I want the 5th and 6th digit. Can Oracle do this? > > If the 5th and 6th digit have meaning on their own, then they should > have been modeled as their own field(s) instead of concatenated into a > composite field. > > That said, yes, one can extract them with simple expressions. Several > examples have already been given. I don't see why regular expressions > are necessary, but they will work. SUBSTR should work, unless the data > is more hinky than you've let on. If it's fixed length data, I definitively agree with you. -- -- http://alvaro.es - �lvaro G. Vicario - Burgos, Spain -- Mi sitio sobre programaci�n web: http://borrame.com -- Mi web de humor satinado: http://www.demogracia.com -- |