data loss with parallel enabled pipelined function - oracle

Question

I have a pipelined function that loads data into file.
The following is the code of function.
CREATE OR REPLACE FUNCTION DATA_UNLOAD
( p_source IN SYS_REFCURSOR,
p_filename IN VARCHAR2,
p_directory IN VARCHAR2
) RETURN dump_ntt PIPELINED PARALLEL_ENABLE (PARTITION p_source BY ANY)
AS
TYPE row_ntt IS TABLE OF VARCHAR2(32767);
v_rows row_ntt;
v_file UTL_FILE.FILE_TYPE;
v_buffer VARCHAR2(32767);
v_sid VARCHAR(255);
v_name VARCHAR2(255);
v_lines PLS_INTEGER := 0;
v_start_dttm TIMESTAMP WITH TIME ZONE:= SYSTIMESTAMP;
v_end_dttm TIMESTAMP WITH TIME ZONE;
c_eol CONSTANT VARCHAR2(1) := CHR(10);
c_eollen CONSTANT PLS_INTEGER := LENGTH(c_eol);
c_maxline CONSTANT PLS_INTEGER := 32767;
BEGIN
--v_sid := lpad(sys_context('USERENV', 'sid'), 10, '0');
v_name:=p_filename;
LOOP
if utl_file.is_open(v_file)
then
utl_file.fclose(v_file);
end if;
v_file := UTL_FILE.FOPEN(p_directory, v_name, 'A', c_maxline);
FETCH p_source BULK COLLECT INTO v_rows LIMIT 100;
FOR i IN 1 .. v_rows.COUNT LOOP
IF LENGTH(v_buffer) + c_eollen + LENGTH(v_rows(i)) <= c_maxline THEN
v_buffer := v_buffer || c_eol || v_rows(i);
ELSE
IF v_buffer IS NOT NULL THEN
UTL_FILE.PUT_LINE(v_file, v_buffer);
END IF;
v_buffer := v_rows(i);
END IF;
END LOOP;
v_lines := v_lines + v_rows.COUNT;
EXIT WHEN p_source%NOTFOUND;
END LOOP;
CLOSE p_source;
UTL_FILE.PUT_LINE(v_file, v_buffer);
UTL_FILE.FCLOSE(v_file);
v_end_dttm := SYSTIMESTAMP;
--PIPE ROW (dump_ot(v_name, p_directory, v_lines, v_sid, v_start_dttm, v_end_dttm));
--RETURN ;
END;
i call the function this way.
SELECT * from table(DATA_UNLOAD(
CURSOR(select /*+ PARALLEL */ a || b || c from sample_table),
'sample.txt',
'99_DIR'));
a real life select that i pass as a parameter to function returns 30000 rows, but when i use the function to load the result into a file some rows are lost. During the execution with PARALLEL hint there are 24 parallel sessions, and i dont want to make it less. My guess is that the problem is in parallel execution, because when i dont use PARALLEL hint no data is lost. Can anyone suggest something to get rid of that problem without removing the hint?

Answer 1

Even though you are creating sample.txt with Append mode - you have 24 parallel sessions each writing to it. I always use unique filenames by appending the SID to your variable:
SELECT sid INTO v_sid FROM v$mystat WHERE ROWNUM = 1;
v_name := p_filename || '_' || v_sid || '.dat';
Depending on the # of parallel sessions you should 1 to many files with the format sample_nnnn.txt where nnnn is the SID number.