Link to home
Start Free TrialLog in
Avatar of jknj72

asked on

Oracle procedure logic - string manipulation

I have a view where I have to do a little string manipulation. I have a text field where I just want to bring back the text in the field after the last pipe | . For instance '~(`|¤|football)~'
I just want to bring back football. There is commonality to the data because they all start with '~( and end with )~'. Also, the data that I need will always be after the last pipe |.
Another example would be '~(`|¤|{link:10098}|Mobile, Alabama,)~' would return
Mobile, Alabama,

Please let me know if you have any questions.

Thanks everyone
Avatar of Muhammad Ahmad Imran
Muhammad Ahmad Imran
Flag of United Kingdom of Great Britain and Northern Ireland image

select substr(yrcol,instr(yrcol,'|',-1),instr(yrcol,')')-2) from yrtab
Avatar of jknj72


this            '~(`|¤|{link:10098}|Mobile, Alabama,)~'
returns       |Mobile, Alabama,)~

I think I still need a bit of tweaking
Avatar of jknj72


Really could use some help here........
In MS SQL syntax it will be somethingn like

SET @Str = '~(`|¤|{link:10098}|Mobile, Alabama,)~'
--SET @Str =  '~(`|¤|football)~'

	,REPLACE(RIGHT(@Str, CHARINDEX('|', REVERSE(@Str))- 1), ')~', '')

Open in new window

Dont have Oracle installed otherwise I would modify it
Avatar of Sean Stuber
Avatar of jknj72


I am getting an error executing ; regexp_substr(str,'.*\|(.+)\)~$',1,1,null,1)

  Execution (28: 8): ORA-00939: too many arguments for function

It seems that there are 6 arguments in your example and looking online I see examples that use are 4(?)

Any ideas?
Would you be kind enough to explain the pattern match and what each of the remaining four parameters are? I typically think of the regexp_substr function as having a maximum of five parameters, the string expression, the pattern, the start position, the occurrence, and a match parameter but you're showing six. Great job, by the way!
It worked for me.
SQL> select * from txt;
~(`|¤|{link:10098}|Mobile, Alabama,)~

SQL> select regexp_substr(text,'.*\|(.+)\)~$',1,1,null,1)
  2  from txt;
Mobile, Alabama,
The last parameter (subexpression) is for 11gR2.  If you have a lower version you'll get the error you see.

the parameters are...

str,   - the string you are searcing
'.*\|(.+)\)~$'  - the regular expression describing what we are looking for
,1  - starting at the first character
,1 - find the first instance of the expression
,null - no special options like case insensitive or multi-line
,1 - return only the first subexpression
Avatar of Sean Stuber
Sean Stuber

Link to home
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of jknj72


Nicely done sdstuber!!  Thank you very much!!
Avatar of jknj72


great work
Thanks so much for the explanation. We recently upgraded to 11gR2, which was why it worked, but was looking at 11gR1 documentation. Regular expressions aren't very intuitive, but extremely powerful when you get them right :-)
Avatar of jknj72


I think I may have screwed up. Its not working 100%, I will post another question related to this one
Avatar of jknj72


please see ID: 27979957
This query takes all characters before the '~ (whether or not they exist) and concatenates them with the function sdstuber provided against all of the characters between the '~ and ~' and concatenates them with all of the characters after the ~'  (whether or not they exist), adding a space between the three parts then trimming any outside spaces -

select trim(substr(text,instr(text,'~'||chr(39)) + 2)||' '||
REPLACE(SUBSTR(REGEXP_SUBSTR(text, '~[^~]+~'), INSTR(REGEXP_SUBSTR(text, '~[^~]+~'), '|', -1) + 1), ')~')||' '||
substr(text,instr(text,'~'||chr(39)) + 2)) mystring
from txt;

Mobile, Alabama,
blahblahblah football blahblahblah
I meant to post this to your other question, which I have now done.