Multiple rows into a single row.

Question

Table structure:
CREATE TABLE JOB_ERROR_MESSAGE_TMP
ERROR_LOG VARCHAR2(1000 BYTE)
This table gets populated with multiple records (approx 20) from the log file of my application.
Example of Data (12 rows below):
ERROR_LOG
XXXXXXXX(Error)-ORA-01756-ZZZZZZZZZZZZZ-quoted string not properly terminated
line1 of error1
line2 of error1
line3 of error1
XXXXXXXX(Error)-DBS-0000-ZZZZZZZZZZZZZ-SOME ERROR
line1 of error2
line2 of error2
XXXXXXXX(Error)-DBS-1111-ZZZZZZZZZZZZZ-SOME NEW ERROR
line1 of error3
line2 of error3
line2 of error4
line4 of error3
I am interested in only the first error. That is: all the rows until the second occurrence of word "(ERROR)"
Desired output:
ORA-01756,quoted string not properly terminated,line1 of error1,line2 of error1,line3 of error1
That is:
The 9 digit code which exists after the 1st encounter of word '(Error)'||','||trim(First line after error)||','||trim(second line after error)||','||
trim of 3rd line <all the lines untill the line containing '(ERROR)' is encountered>
Any help on this is appreciated.
Thanks in advance
Coolguy

Answer 1

Example...
SQL> ed
Wrote file afiedt.buf
  1  with t as (select 1 as ord, 'XXXXXXXX(Error)-ORA-01756-ZZZZZZZZZZZZZ-quoted string not properly terminated' as txt from dual union all
  2             select 2, 'line1 of error1' from dual union all
  3             select 3, 'line2 of error1' from dual union all
  4             select 4, 'line3 of error1' from dual union all
  5             select 5, 'XXXXXXXX(Error)-DBS-0000-ZZZZZZZZZZZZZ-SOME ERROR' from dual union all
  6             select 6, 'line1 of error2' from dual union all
  7             select 7, 'line2 of error2' from dual union all
  8             select 8, 'XXXXXXXX(Error)-DBS-1111-ZZZZZZZZZZZZZ-SOME NEW ERROR' from dual union all
  9             select 9, 'line1 of error3' from dual union all
10             select 10, 'line2 of error3' from dual union all
11             select 11, 'line2 of error4' from dual union all
12             select 12, 'line4 of error3' from dual)
13  -- end of test data
14  select txt
15  from (
16  select ord, txt, starter, lag(starter) over (order by ord) as prev
17  from (
18  select ord, txt, decode(instr(txt,'(Error)'),0,null,1) as starter
19  from t
20  order by ord
21  )
22  )
23  where coalesce(starter, prev) = 1
24* order by ord
SQL> /
TXT
XXXXXXXX(Error)-ORA-01756-ZZZZZZZZZZZZZ-quoted string not properly terminated
line1 of error1
XXXXXXXX(Error)-DBS-0000-ZZZZZZZZZZZZZ-SOME ERROR
line1 of error2
XXXXXXXX(Error)-DBS-1111-ZZZZZZZZZZZZZ-SOME NEW ERROR
line1 of error3
6 rows selected.
SQL>Your problem is going to be that you haven't got anything in your table which specifies the order of the rows. As such you have an unordered set of data in a table and there is nothing to say which rows of data are associated with other rows. If you want it to work you're going to have to update your table to include a sequence number or a date/timestamp.