oracle 管道化表函数(Pipelined Table)

    在实际的应用中,为了让PL/SQL 函数返回数据的多个行,必须通过返回一个 REF CURSOR 或一个数据集合来完成。REF CURSOR 的这种情况局限于可以从查询中选择的数据,而整个集合在可以返回前,必须进行具体化。 9i 通过引入的管道化表函数纠正了后一种情况。表函数是返回整个行的集(通常作为一个集合)的函数,可以直接从 SQL 语句中进行查询,就好像它是一个真正的数据库表一样。管道化表函数与之相似,但是它像在构建时一样返回数据,而不是一次全部返回。管道化表函数更加有效,因为数据可以尽可能快地返回。

    管道化表函数必须返回一个集合。在函数中,PIPE ROW 语句被用来返回该集合的单个元素,该函数必须以一个空的 RETURN 语句结束,以表明它已经完成。一旦我们创建了上述函数,我们就可以使用 TABLE 操作符从 SQL 查询中调用它。

例1:
Type Definition
CREATE OR REPLACE TYPE COLOR_HEX_CODE AS OBJECT
(
  -- Attributes
  colorName Varchar2(10),
  colorCode varchar2(10)
  
  -- Member functions and procedures
  --member procedure <ProcedureName>(<Parameter> <Datatype>)
);

Create Table Type
CREATE OR REPLACE TYPE COLOR_TYPE AS TABLE OF COLOR_HEX_CODE;

Create Table Function
CREATE OR REPLACE FUNCTION F_PIPELINE_TEST RETURN COLOR_TYPE PIPELINED
AS
BEGIN
    PIPE ROW(COLOR_HEX_CODE('black','#000000'));
    PIPE ROW(COLOR_HEX_CODE('white','#FFFFFF'));
    PIPE ROW(COLOR_HEX_CODE('red','#FF0000'));
    PIPE ROW(COLOR_HEX_CODE('green','#00FF00'));
    PIPE ROW(COLOR_HEX_CODE('blue','#0000FF'));
    RETURN;
END;

测试:
select * from table( f_pipeline_test );

结果:
COLORNAME  COLORCODE
black   #000000
white   #FFFFFF
red   #FF0000
green   #00FF00
blue   #0000FF

例2:实现split函数功能
CREATE OR REPLACE TYPE RESOLVE_STR IS TABLE OF VARCHAR2 (4000);

CREATE OR REPLACE FUNCTION F_SPLIT
    (
       P_STR       IN   VARCHAR2,
       P_DELIMITER IN   varchar2
     ) RETURN RESOLVE_STR PIPELINED
AS
    j     INT := 0;
    i     INT := 1;
    len   INT := 0;
    len1  INT := 0;
    tmp   VARCHAR2 (4000);
    v_str VARCHAR2 (4000);
BEGIN
    v_str := TRIM(BOTH P_DELIMITER FROM P_STR);--去掉前后的分隔符
    len := LENGTH (v_str);
    len1 := LENGTH(P_DELIMITER);

    WHILE j < len
      LOOP
          j := INSTR (v_str, P_DELIMITER, i);--寻分隔符位置

          IF j = 0 THEN--没有了
              tmp := SUBSTR (v_str, i);
              PIPE ROW (tmp);
              EXIT;--结束
          ELSE
              tmp := SUBSTR (v_str, i, j - i);
              i := j + len1;
              PIPE ROW (tmp);
          END IF;
      END LOOP;

    RETURN;
END;

测试:
SELECT * FROM TABLE (F_SPLIT('-12-02-ab-cd-a0-ef-', '-'));

例3:产生6个0..49的随机数
First the quick-and-dirty solution without a pipelined function
select r from (
       select r from (
              select rownum r from all_objects where rownum < 50
                     ) order by dbms_random.VALUE
              )where rownum <= 6;

pipelined function
CREATE OR REPLACE TYPE ARRAY AS TABLE OF NUMBER;

CREATE OR REPLACE FUNCTION gen_numbers(n IN NUMBER DEFAULT NULL)
RETURN ARRAY PIPELINED
AS
  BEGIN
    FOR i IN 1 .. nvl(n,9999)
    LOOP
    PIPE ROW(i);
    END LOOP;
  RETURN;
END;

测试:
select * from (
  select *
    from (select * from table(gen_numbers(49)))  order by dbms_random.random
  )
where rownum <= 6

generating that range of dates
select to_date('2009-5-10','yyyy-mm-dd') + column_value-1 from TABLE(gen_numbers(15));

Note the name of the column we used: COLUMN_VALUE. That is the default name for the column coming back from the PIPELINED function.
another example
CREATE OR REPLACE TYPE date_array AS TABLE OF DATE;

CREATE OR REPLACE FUNCTION date_table(sdate DATE, edate DATE)
RETURN date_array PIPELINED AS
BEGIN
  FOR i IN 0 .. (edate - sdate)
  LOOP
    PIPE ROW(sdate + i);
  END LOOP;
  RETURN;
END date_table;

你可能感兴趣的:(oracle,sql,F#,J#)