Convert comma delimed values to distinct records

Question

Hello,
i wrote the following code to get the data from comma delimited row into distinct values to another table. say for example my data into column looks like this.
long_string column data looks like this:
11,12,14,16,13,18,17
11,12,13,15,16,17,18,19,20,30
11,20,30,40,67,90,87,65,76,43,23,56,78,90,12,22,34 - which doesn't have a comma at the end of the data.
i need to convert this into distinct record for each rows data.
i used mod function...
can some one help me.
DECLARE
cursor c_1 is select LONG_STRING from TEST;
rec c_1%rowtype;
temp_VAR varchar2(100);
size_of_VAR number;
number_of_VAR number;
i number;
cur_VAR varchar2(3);
begin
open c_1;
loop
fetch c_1 into rec;
exit when c_1%notfound;
temp_VAR := rec.LONG_STRING;
size_of_VAR := length(temp_VAR);
number_of_VAR := size_of_VAR/4;
if (mod (size_of_VAR,4) = 0) then
for i in 1..number_of_VAR loop
cur_VAR:=substr(temp_VAR,1,3);
INSERT INTO TEST_ST (LONG_ST_NUM)
VALUES ( cur_VAR);
temp_VAR:= substr(temp_VAR,5);
end loop;
else
dbms_output.put_line('Rules problem for LONG_STRING: ' || rec.LONG_STRING);
end if;
end loop;
close c_1;
end;
Thanks...

Answer 1

Ah, I see, I took my requirement from the later posters as it wasn't clear from OP. It's not the aggregation problem that Listagg solves, but the inverse, disaggregation. I don't know a direct equivalent SQL function, but not to worry, it seems my weekend effort on another thread wasn't wasted because I can see a very simple variation on one of the solutions there that will solve this one very efficiently. It uses the model clause to iterate over the commas without any regular expression functions, XML, tree walks etc. to slow it down, and this approach should work for many similar problems. I'll add an id column so we know where the rows came from and a subid for ordering (you can drop them if you like), and an extra test record with no comma.
CREATE TABLE test_otn  (id INTEGER, text VARCHAR2(4000));
INSERT INTO test_otn VALUES (1, '11,12,14,16,13,18,17');
INSERT INTO test_otn VALUES (2, '11,12,13,15,16,17,18,19,20,30');
INSERT INTO test_otn VALUES (3, '99');
INSERT INTO test_otn VALUES (4, '11,20,30,40,67,90,87,65,76,43,23,56,78,90,12,22,34');
SELECT id, numid, num
  FROM test_otn
MODEL
    PARTITION BY (id)
    DIMENSION BY (0 i)
    MEASURES (0 compos, 0 nxtpos, text, 0 numid, CAST ('' AS VARCHAR2(5)) num)
    RULES
    ITERATE (100) UNTIL compos[0] = 0
       nxtpos[0] = Instr (text[0], ',', compos[0] + 1, 1),
       num[ITERATION_NUMBER] = CASE WHEN nxtpos[0] = 0 THEN Substr (text[0], compos[0]+1)
                                    ELSE Substr (text[0], compos[0] + 1, nxtpos[0] - compos[0] - 1) END,
       numid[ITERATION_NUMBER] = ITERATION_NUMBER + 1,
       compos[0] = nxtpos[0]
ORDER BY 1, 2
  ID      NUMID NUM
   1          1 11
              2 12
              3 14
              4 16
              5 13
              6 18
              7 17
   2          1 11
              2 12
              3 13
              4 15
              5 16
              6 17
              7 18
              8 19
              9 20
             10 30
   3          1 99
   4          1 11
              2 20
              3 30
              4 40
              5 67
              6 90
              7 87
              8 65
              9 76
             10 43
             11 23
             12 56
             13 78
             14 90
             15 12
             16 22
             17 34
35 rows selected.