SQL - Combining information with similar primary key/ID - sql

Question

I know this question was asked before but the solution given for that question did not work for me. I couldn't write a comment stating my problem as I needed 50 reputation.
Details:
Oracle version 10g
Solution Provided:
select book_id,
'Total Hours of Loan: ' || to_char(sum(hours_of_loan)) || chr(10) ||
lpad('-', 30, '-') || chr(10) ||
xmlagg(xmlelement(x, descr, chr(10) || lpad('-', 30, '-') || chr(10)).extract('//text()')
order by date_of_loan) as book_description
from (
select book_id, date_of_loan, hours_of_loan,
'Written by: ' || checked_by || chr(10) ||
'Date of Loan : ' || to_char(date_of_loan, 'mm/dd/yy') || chr(10) ||
'Hour(s) of Loan: ' || to_char(hours_of_loan) || chr(10) ||
'Comments: ' || comments || chr(10) ||
'Student: ' || student as descr
from student_book
)
group by book_id
order by book_id
;
I tried creating a dummy table with fake data to test out the logic to see if it works for me and it does. But when I used the query for my real data, it prompted out an error that says:
ORA-22813: operand value exceeds system limits
ERROR prompted
After reading about it, I found that it was because my data was too large.
I tried creating table and a query to call from it
DECLARE
sSUM_TEXT CLOB;
sBOOK_ID VARCHAR2(10):= '1';
nCount NUMBER:=1;
BEGIN
FOR rRECORDRec IN(SELECT BOOK_ID, CHECKED_BY, DATE_OF_lOAN, HOURS_OF_LOAN, COMMENTS, STUDENT
FROM STUDENT_BOOK
WHERE BOOK_ID = '1')
LOOP
IF nCount != 1 THEN
sSUM_TEXT := sSUM_TEXT||chr(10)||
'Written by: ' || rRECORDRec.checked_by || chr(10) ||
'Date of Loan : ' || rRECORDRec.date_of_loan || chr(10) ||
'Hour(s) of Loan: ' || rRECORDRec.hours_of_loan || chr(10) ||
'Comments: ' || rRECORDRec.comments || chr(10) ||
'Student: ' || rRECORDRec.student || chr(10) ||
'--------------------------------';
ELSE
sSUM_TEXT := 'Written by: ' || rRECORDRec.checked_by || chr(10) ||
'Date of Loan : ' || rRECORDRec.date_of_loan || chr(10) ||
'Hour(s) of Loan: ' || rRECORDRec.hours_of_loan || chr(10) ||
'Comments: ' || rRECORDRec.comments || chr(10) ||
'Student: ' || rRECORDRec.student || chr(10) ||
'--------------------------------';
END IF;
nCount := nCount + 1;
END LOOP;
INSERT INTO COMMENTS_SUMMARY(BOOK_ID,SUM_TEXT)
VALUES (sBOOK_ID,sSUM_TEXT);
COMMIT;
END;
This did not really solve my problem but it combined all the "COMMENTS" which had that specific BOOK_ID but without format
Tried Solution
select book_id,
'Total Hours of Loan: ' || to_char(sum(hours_of_loan)) || chr(10) ||
lpad('-', 30, '-') || chr(10) ||
xmlagg(xmlelement(x, descr, chr(10) || lpad('-', 30, '-') || chr(10)).extract('//text()')
order by date_of_loan) as book_description
from (
select book_id, date_of_loan, hours_of_loan,
to_clob('Written by: ' || checked_by || chr(10) ||
'Date of Loan : ' || date_of_loan || chr(10) ||
'Hour(s) of Loan: ' || hours_of_loan || chr(10) ||
'Comments: ' || comments || chr(10) ||
'Student: ' || student) as descr
from student_book
)
group by book_id
order by book_id;
The same error prompted out
2nd SELECT Statement
I tried running the query separately. I ran the 2nd SELECT Statement for my real data and no errors were prompted even without the TO_CLOB
select book_id, date_of_loan, hours_of_loan,
'Written by: ' || checked_by || chr(10) ||
'Date of Loan : ' || date_of_loan || chr(10) ||
'Hour(s) of Loan: ' || hours_of_loan || chr(10) ||
'Comments: ' || comments || chr(10) ||
'Student: ' || student as descr
from student_book
order by book_id
)
--group by book_id
order by book_id;
My Question:
How may I overcome this problem and group the data base on their ID?

Answer 1

There are probably at least a few books for which you have too many loans, so the string built in XMLAGG is longer than 4000 characters. So you have to deal with CLOBs.
You need to add .getclobval() here:
... order by date_of_loan).getclobval() as book_description
It is possible that even the text for one "loan" is already too long (more than 4000 characters). This is a problem for concatenation - if all the inputs are less than 4000 characters Oracle treats them as VARCHAR2 and expects the result to also be VARCHAR2 (no more than 4000 characters). To "force" it to treat everything as CLOBs, it suffices to make the first input a CLOB.
Instead of 'Written by: ' (find it in the code) use
... to_clob('Written by: ')