To validate each part of the string

Question

Hi,
In one string, we have these
ID_A,ID_B,ID_C, ..., ...
I know that there's a way in PL/SQL to divide it into several parts and each part is for one ID. But can I expect that to divide the string into several parts, and then to store all IDs in something like an Array, in PL/SQL? Is this possible, though I know there's no array field type in PL/SQL?
Regards,
huamin

Answer 1

xtender wrote:
You can use my package xt_regexp - http://github.com/xtender/XT_REGEXP
To install execute in this order:     types.sql,     xt_regexp.jsp,     xt_regexp.pck
Example:
with t as (
select 'ID_A,ID_B,ID_C,' str from dual
union all
select 'ID_A,ID_B,ID_C,ID_D' str from dual
select
t.*,
row_number() over (partition by t.str order by t.str) substr_number,
substring.column_value substring
from t,
table(xt_regexp.split(t.str,',')) substring
That's nice, but you may want to benchmark that against an approach that doesn't make a bunch of external calls (for performance)...
with t as
  select 1 as id,'ID_A,ID_B,ID_C' str from dual
    union all
  select 2 as id, 'ID_A,ID_B,ID_C,ID_D' str from dual
select
  t.*,
  regexp_substr (t.str, '[^,]+', 1, s.column_value) as split
from
  t,
  table( cast (multiset (select level from dual connect by level <= length(regexp_replace(t.str, '[^,]+'))  + 1) as sys.odcinumberlist)) s
     ID STR                 SPLIT
      1 ID_A,ID_B,ID_C      ID_A
      1 ID_A,ID_B,ID_C      ID_B
      1 ID_A,ID_B,ID_C      ID_C
      2 ID_A,ID_B,ID_C,ID_D ID_A
      2 ID_A,ID_B,ID_C,ID_D ID_B
      2 ID_A,ID_B,ID_C,ID_D ID_C
      2 ID_A,ID_B,ID_C,ID_D ID_D
7 rows selected.
Elapsed: 00:00:00.02
TUBBY_TUBBZ?For example.