Trouble with script to break out csv variables

Question

Using the following script to break down a csv file.
CREATE OR REPLACE PROCEDURE s1_prc is
cursor sa_cur is
with t as (
select sa_record str from sa_records
-- where rownum < 2
select regexp_substr(str,'[^' || CHR (9) || ']+',1,1) c1,
regexp_substr(str,'[^' || CHR (9) || ']+',1,2) c2,
regexp_substr(str,'[^' || CHR (9) || ']+',1,3) c3,
regexp_substr(str,'[^' || CHR (9) || ']+',1,4) c4,
regexp_substr(str,'[^' || CHR (9) || ']+',1,5) c5,
regexp_substr(str,'[^' || CHR (9) || ']+',1,6) c6,
regexp_substr(str,'[^' || CHR (9) || ']+',1,7) c7,
regexp_substr(str,'[^' || CHR (9) || ']+',1,8) c8,
regexp_substr(str,'[^' || CHR (9) || ']+',1,9) c9,
regexp_substr(str,'[^' || CHR (9) || ']+',1,10) c10,
regexp_substr(str,'[^' || CHR (9) || ']+',1,11) c11,
regexp_substr(str,'[^' || CHR (9) || ']+',1,12) c12,
regexp_substr(str,'[^' || CHR (9) || ']+',1,13) c13,
regexp_substr(str,'[^' || CHR (9) || ']+',1,14) c14,
regexp_substr(str,'[^' || CHR (9) || ']+',1,15) c15,
regexp_substr(str,'[^' || CHR (9) || ']+',1,16) c16,
regexp_substr(str,'[^' || CHR (9) || ']+',1,17) c17,
regexp_substr(str,'[^' || CHR (9) || ']+',1,18) c18,
regexp_substr(str,'[^' || CHR (9) || ']+',1,19) c19,
regexp_substr(str,'[^' || CHR (9) || ']+',1,20) c20,
regexp_substr(str,'[^' || CHR (9) || ']+',1,21) c21,
regexp_substr(str,'[^' || CHR (9) || ']+',1,22) c22,
regexp_substr(str,'[^' || CHR (9) || ']+',1,23) c23,
regexp_substr(str,'[^' || CHR (9) || ']+',1,24) c24,
regexp_substr(str,'[^' || CHR (9) || ']+',1,25) c25
from t;
BEGIN
delete from reciprocal_records;
commit;
for i in sa_cur loop
DBMS_OUTPUT.PUT_LINE ('1 '||i.c1);
DBMS_OUTPUT.PUT_LINE ('2 '||i.c2);
DBMS_OUTPUT.PUT_LINE ('3 '||i.c3);
DBMS_OUTPUT.PUT_LINE ('4 '||i.c4);
DBMS_OUTPUT.PUT_LINE ('5 '||i.c5);
DBMS_OUTPUT.PUT_LINE ('6 '||i.c6);
DBMS_OUTPUT.PUT_LINE ('7 '||i.c7);
DBMS_OUTPUT.PUT_LINE ('8 '||i.c8);
DBMS_OUTPUT.PUT_LINE ('9 '||i.c9);
DBMS_OUTPUT.PUT_LINE ('10 '||i.c10);
DBMS_OUTPUT.PUT_LINE ('11 '||i.c11);
DBMS_OUTPUT.PUT_LINE ('12 '||i.c12);
DBMS_OUTPUT.PUT_LINE ('13 '||i.c13);
DBMS_OUTPUT.PUT_LINE ('14 '||i.c14);
DBMS_OUTPUT.PUT_LINE ('15 '||i.c15);
DBMS_OUTPUT.PUT_LINE ('16 '||i.c16);
DBMS_OUTPUT.PUT_LINE ('17 '||i.c17);
DBMS_OUTPUT.PUT_LINE ('18 '||i.c18);
DBMS_OUTPUT.PUT_LINE ('19 '||i.c19);
DBMS_OUTPUT.PUT_LINE ('20 '||i.c20);
DBMS_OUTPUT.PUT_LINE ('21 '||i.c21);
DBMS_OUTPUT.PUT_LINE ('22 '||i.c22);
DBMS_OUTPUT.PUT_LINE ('23 '||i.c23);
DBMS_OUTPUT.PUT_LINE ('24 '||i.c24);
DBMS_OUTPUT.PUT_LINE ('25 '||i.c25);
insert into in_records (IN_SEQ_NO,
IN_TRANS_TYPE, -- 1
IN_SUBMIT_CLUB, -- 2
IN_PROC_ID, -- 3
IN_SERV_CLUB, -- 4
IN_SERVICE_DATE, -- 5
IN_CALL_ID, -- 6
IN_RAP_CALL_ID, -- 7
IN_FACILITY_CODE,-- 8
IN_TROUBLE_CODE, -- 9
IN_DUTY_FLAG, -- 10
IN_CLUB_CODE, -- 11
IN_MEMBER_NO, -- 12
IN_MEMBER_TYPE, -- 13
IN_EXPIRY_DATE, -- 14
IN_LAST_NAME, -- 15
IN_FIRST_NAME, -- 16
IN_EDS_VIN_NO, -- 17
IN_ODOMETER, -- 18
IN_TRAVEL, -- 19
IN_MILES_DEST, -- 20
IN_TAX_AMT, -- 21
IN_TOTAL_AMT, -- 22
IN_DISPUTE_CODE, -- 23
IN_ERR_CODE, -- 24
IN_COST) -- 25
values (1, i.c1, i.c2, i.c3, i.c4, i.c5, i.c6, i.c7, i.c8, i.c9, i.c10, i.c11, i.c12, i.c13, i.c14, i.c15, i.c16, i.c17, i.c18, i.c19, i.c20, i.c21, i.c22, i.c23, i.c24, i.c25);
commit;
end loop;
commit;
end;
The problem I'm having is if there are two null characters together, it seem that the script is ignoring the first null character.
Example:
ACP^     212^     HOC^     212^     08292012^     1069^     N^     1603^     T1^          097^     4290972957271019^     PLUS^     03312013^     GAGLIO^     MARC^ ^          N^     010^     000^     000149^     002746^     ^     ^     PLS/02567/:TEL/00030/TELECOMMUNICATIONS CHARGE
After the first name, Marc, there are two (CHR 9) characters, also after 002746 there are two (CHR 9) characters. So what happens is the last field 'PLS/02567...' ends up as the 23rd field, not as the 25th field.
Hope this clear. Any suggestions are welcome.
Thanks
Paul

Answer 1

Hi, Paul,
To find the item before the n-th CHR(9), even if that item is NULL:
RTRIM ( REGEXP_SUBSTR ( str
                , '[^' || CHR (9) || ']*(' || CHR  (9) || '|$)'
                , 1
                , n
      , CHR (9)
      )The expression above will work in Oracle 10.1 and higher.
Starting in Oracle 11.1, you can do this without RTRIM, because REGEXP_SUBSTR can now be told to return only part of the matching pattern.
Edited by: Frank Kulash on Oct 15, 2012 1:36 PM