oracel 拆分字符串

CREATE OR REPLACE TYPE str_split IS TABLE OF VARCHAR2 (4000);
CREATE OR REPLACE FUNCTION splitstr(p_string IN VARCHAR2, p_delimiter IN VARCHAR2)
    RETURN str_split 
    PIPELINED
AS
    v_length   NUMBER := LENGTH(p_string);
    v_start    NUMBER := 1;
    v_index    NUMBER;
BEGIN
    WHILE(v_start <= v_length)
    LOOP
        v_index := INSTR(p_string, p_delimiter, v_start);

        IF v_index = 0
        THEN
            PIPE ROW(SUBSTR(p_string, v_start));
            v_start := v_length + 1;
        ELSE
            PIPE ROW(SUBSTR(p_string, v_start, v_index - v_start));
            v_start := v_index + 1;
        END IF;
    END LOOP;

    RETURN;
END splitstr;

创建完毕后，我们来测试一下，例如执行如下SQL：

select * from table(splitstr('Hello,Cnblogs!',','));

其输出结果为一个两行的表，如下图：

将行转为列显示：

select a.column_value v1,b.column_value v2 from 
(select * from (select rownum rn,t.* from table(splitstr('Hello,Cnblogs!',',')) t)) a,
(select * from (select rownum rn,t.* from table(splitstr('Hello,Cnblogs!',',')) t)) b
where a.rn=1 and b.rn=2

如图：

作者：李敬然（Gnie） 出处：{GnieTech} （http://www.cnblogs.com/gnielee/）

 

 

方法二、

select regexp_substr('Hello,Cnblogs!','[^,]+',1,level) from dual connect by level <= length('Hello,Cnblogs!')-length(REPLACE('Hello,Cnblogs!',',',''))+1;