How to convert CLOB to varchar2 whose length is more than 4000

Question

Hi ,
I have to retrive the CLOB data into varchar2 and spilt those comma seperated values into rows
but i am getting following error
" ORA-06502: PL/SQL: numeric or value error: character string buffer too small "
the proc i used is like following :
: create table test_clob (grp_id CLOB, id number)
Create or replace proc test_clob(p_id number )
Is
V_clob CLOB;
V_str varchar2(4000);
V_main varchar2(30000);
TYPE t_clob IS REF CURSOR;
cur_clob t_clob;
Begin
Select grp_id
Into v_str
From test_clob
Where id= p_id;
---converting column data in rows as I have to return the cursor as output
V_main:= ' select grp_id from ( WITH t AS
(SELECT REGEXP_SUBSTR(''' || v_str ||
''', ''[^,]+'', 1, LEVEL) txt
FROM DUAL
CONNECT BY LEVEL <=
LENGTH(''' || v_str ||
LENGTH(REPLACE(''' || v_str ||
''', '','')) + 1)
SELECT REGEXP_SUBSTR(trim(txt), ''[^\,]+'', 1, 1) grp_id
FROM t )';
open cur_clob for ' select rtrim(xmlagg(xmlelement(e, grp_id || '','')).extract (''//text()'').getclobval (),'','') grp_id
from ( '|| v_main||' ) ';
loop
fetch cur_clob
into V_clob;
exit when cur_clob %notfound;
end loop;
insert into test_clob
values (p_id, v_clob);
commit;
End;
Please help its very urgent

Answer 1

957624 wrote:
Hi ,
I have to retrive the CLOB data into varchar2 and spilt those comma seperated values into rows
but i am getting following error
" ORA-06502: PL/SQL: numeric or value error: character string buffer too small "
the proc i used is like following :
: create table test_clob (grp_id CLOB, id number)
Create or replace proc test_clob(p_id number )
Is
V_clob CLOB;
V_str varchar2(4000);
V_main varchar2(30000);
TYPE t_clob IS REF CURSOR;
cur_clob t_clob;
Begin
Select grp_id
Into v_str
From test_clob
Where id= p_id;
---converting column data in rows as I have to return the cursor as output
V_main:= ' select grp_id from ( WITH t AS
(SELECT REGEXP_SUBSTR(''' || v_str ||
''', ''[^,]+'', 1, LEVEL) txt
FROM DUAL
CONNECT BY LEVEL <=
LENGTH(''' || v_str ||
LENGTH(REPLACE(''' || v_str ||
''', '','')) + 1)
SELECT REGEXP_SUBSTR(trim(txt), ''[^\,]+'', 1, 1) grp_id
FROM t )';
open cur_clob for ' select rtrim(xmlagg(xmlelement(e, grp_id || '','')).extract (''//text()'').getclobval (),'','') grp_id
from ( '|| v_main||' ) ';
loop
fetch cur_clob
into V_clob;
exit when cur_clob %notfound;
end loop;
insert into test_clob
values (p_id, v_clob);
commit;
End;
Please help its very urgentNo, nothing is urgent here and it is very rude to suggest it is, as well as it being a breach of the terms of use for the forums. Urgent issues relate to commercially live systems, and for those you need to raise a support request with Oracle Support. The forums are manned by volunteers with their own jobs to do, so suggesting they drop everything to help you urgently is very rude. Also, other people asking questions would like answers to theirs as soon as possible too, so suggesting your question is more urgent than theirs is also rude to those people. You will find that suggesting your question is urgent will actually prevent people from helping you on the forums as many people will simply choose to ignore you for your rudeness (either that or they'll be rude back)
Please read {message:id=9360002} to learn how to ask a question properly.
Now, in answer to your question...
What is it you are really trying to do?
If you have comma seperated data, why it is being stored like that in a CLOB? That seems like bad design for storing data.
If the data is coming from a file, then you could consider using External Tables (or SQL*Loader) to load the comma seperated data in a structured way.
If you really have data in a clob in comma seperated format, what are you wanting to do with it exactly i.e. what determines where the data should be split?
Your code isn't clear, but it looks like, on the one hand you are splitting comma seperated data out of a string, and on the other hand you're turning it into XML to get a clob of data back.
Please be clear by providing some example data and expected output, along with your database version as explained in the FAQ post I linked to above.