Oracle 管道化表函数(Pipelined Table)[转载]
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
1 CREATE OR REPLACE TYPE COLOR_HEX_CODE AS OBJECT
2 (
3 -- Attributes
4 colorName Varchar2(10),
5 colorCode varchar2(10)
6
7 -- Member functions and procedures
8 --member procedure <ProcedureName>(<Parameter> <Datatype>)
9 );
Create Table Type
CREATE OR REPLACE TYPE COLOR_TYPE AS TABLE OF COLOR_HEX_CODE;
Create Table Function
1 CREATE OR REPLACE FUNCTION F_PIPELINE_TEST RETURN COLOR_TYPE PIPELINED
2 AS
3 BEGIN
4 PIPE ROW(COLOR_HEX_CODE('black','#000000'));
5 PIPE ROW(COLOR_HEX_CODE('white','#FFFFFF'));
6 PIPE ROW(COLOR_HEX_CODE('red','#FF0000'));
7 PIPE ROW(COLOR_HEX_CODE('green','#00FF00'));
8 PIPE ROW(COLOR_HEX_CODE('blue','#0000FF'));
9 RETURN;
10 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);
1 CREATE OR REPLACE FUNCTION F_SPLIT
2 (
3 P_STR IN VARCHAR2,
4 P_DELIMITER IN varchar2
5 ) RETURN RESOLVE_STR PIPELINED
6 AS
7 j INT := 0;
8 i INT := 1;
9 len INT := 0;
10 len1 INT := 0;
11 tmp VARCHAR2 (4000);
12 v_str VARCHAR2 (4000);
13 BEGIN
14 v_str := TRIM(BOTH P_DELIMITER FROM P_STR);--去掉前后的分隔符
15 len := LENGTH (v_str);
16 len1 := LENGTH(P_DELIMITER);
17
18 WHILE j < len
19 LOOP
20 j := INSTR (v_str, P_DELIMITER, i);--寻分隔符位置
21
22 IF j = 0 THEN--没有了
23 tmp := SUBSTR (v_str, i);
24 PIPE ROW (tmp);
25 EXIT;--结束
26 ELSE
27 tmp := SUBSTR (v_str, i, j - i);
28 i := j + len1;
29 PIPE ROW (tmp);
30 END IF;
31 END LOOP;
32
33 RETURN;
34 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
1 select r from (
2 select r from (
3 select rownum r from all_objects where rownum < 50
4 ) order by dbms_random.VALUE
5 )where rownum <= 6;
pipelined function
CREATE OR REPLACE TYPE ARRAY AS TABLE OF NUMBER;
1 CREATE OR REPLACE FUNCTION gen_numbers(n IN NUMBER DEFAULT NULL)
2 RETURN ARRAY PIPELINED
3 AS
4 BEGIN
5 FOR i IN 1 .. nvl(n,9999)
6 LOOP
7 PIPE ROW(i);
8 END LOOP;
9 RETURN;
10 END;
测试:
1 select * from (
2 select *
3 from (select * from table(gen_numbers(49))) order by dbms_random.random
4 )
5 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;
1 CREATE OR REPLACE FUNCTION date_table(sdate DATE, edate DATE)
2 RETURN date_array PIPELINED AS
3 BEGIN
4 FOR i IN 0 .. (edate - sdate)
5 LOOP
6 PIPE ROW(sdate + i);
7 END LOOP;
8 RETURN;
9 END date_table;