查询oracle 中逗号分隔字符串中所有值

如果一个字符串中有像逗号或其它符号分隔，你想把它折分成列，如’first field, second field , third field’,

拆成

first field

second field

third field

第一种 用10G开始支持的正则表达式

SELECT REGEXP_SUBSTR (‘first field, second field , third field’, ‘[^,]+’, 1,rownum)
FROM DUAL
CONNECT BY ROWNUM <=
LENGTH (‘first field, second field , third field’) – LENGTH (REPLACE (‘first field, second field , third field’, ‘,’, ”))
+1

————

first field
second field
third field

REGEXP_SUBSTR 函数是把那个串以正则不是以,（逗号）开头的截取，第二个参数是取第几组，rownum伪列序号，connect 循环 ，循环次数为串总长度-去除分隔符后=几个分隔符 +1

第二种用type,function
第一，先创建一个Type

CREATE OR REPLACE TYPE type_split IS TABLE OF VARCHAR2 (4000)

第二，创建函数

create or replace function split(p_list varchar2,p_sep varchar2 := ’,’)
return type_split pipelined
IS
l_idx pls_integer;
v_list varchar2(50) := p_list;
begin
loop
l_idx := instr(v_list,p_sep);
if l_idx > 0 then
pipe row(substr(v_list,1,l_idx-1));
v_list := substr(v_list,l_idx+length(p_sep));
else
pipe row(v_list);
exit;
end if;
end loop;
end split;

第三，调试

select * from table(split(’aaa,bbb,ccc’,’,’))

扩展 regexp_replace

V字段中每个值中字符串以，分隔，如果不是以9开头那组串加‘00’

anbob@NCME>create table testreg(v varchar2(80));
Table created.
anbob@NCME>insert into testreg values(’911,000,12,31′);
1 row created.

anbob@NCME>insert into testreg values(’11911,554000,312,931′);
1 row created.

anbob@NCME>commit;
Commit complete.

anbob@NCME>select ltrim(regexp_replace(‘,’||v,’([,])’,'\100′),’,') newv,v from testreg;

NEWV V
—————————— ——————————
00911,00000,0012,0031 911,000,12,31
0011911,00554000,00312,00931 11911,554000,312,931