Oracle 字串轉數組的函數

create type char_table is table of varchar2(4000);--建立自訂類型指令碼

create or replace function split_string(pi_str in varchar2, pi_separator in varchar2) --建立函數
return char_table
 is
  v_char_table char_table;
  --create type char_table is table of varchar2(4000);--建立自訂類型指令碼
  v_temp varchar2(4000);
  v_element varchar2(4000);
begin
   v_char_table := char_table();
   v_temp := pi_str;
   while instr(v_temp, pi_separator) > 0
    loop
        v_element := substr(v_temp,1,instr(v_temp, pi_separator)-1);
        v_temp := substr(v_temp, instr(v_temp,pi_separator)+ length(pi_separator) , length(v_temp));
        v_char_table.extend;
        v_char_table(v_char_table.count) := v_element;

   end loop;
    v_char_table.extend;
    v_char_table(v_char_table.count) := v_temp;
   return v_char_table;
end split_string;

--測試

select split_string('safs,utjtje,2412q1a,egwsff',',') from dual;

--實際應用

create or replace function initCFJG
  return varchar2
  is
  pragma AUTONOMOUS_TRANSACTION;
guid varchar2(80):=''; --GUID
zjid varchar2(80):='';--主鍵id
lxid varchar2(80):=''; --處理結果類型id
lxidsArr char_table; --字串數組，char_table為自訂類型
begin
     for p in(select lp_id,zfgl_id,case_id,cf_amount from dzjc_xzzf_t_zfgl where lp_id is not null)
      loop
        zjid:=p.zfgl_id;
        select split_string(p.lp_id,',') into lxidsArr from dual;
        for i in 1 .. lxidsArr.count 
        loop  
               --dbms_output.put_line(lxidsArr(i)); 
               select GETGUID into guid from dual;
             
               select cfjd_cflx into lxid from cf_t_cfjd where rownum =1 and cfjd_id=lxidsArr(i);
               if lxid='3'then
                  insert into dzjc_xzzf_t_zfgl(zfgl_id,case_id,lp_id,cf_amount) values (guid,p.case_id,'159_'||lxid,p.cf_amount);
               else
                   insert into dzjc_xzzf_t_zfgl(zfgl_id,case_id,lp_id) values (guid,p.case_id,'159_'||lxid);
               end if;
        end loop;
        delete from dzjc_xzzf_t_zfgl where zfgl_id=zjid;
      end loop;
      commit;
     return '執行完成';
end;