Question

在Oracle SQL Developer中，我使用WITH子句，采用这种（简化）方式：

WITH
foos AS
    SELECT *
    FROM my_table
    WHERE field = 'foo'
bars AS
    SELECT *
    FROM my_table
    WHERE field = 'bar'
SELECT *
FROM foo
INNER JOIN bar
ON foo.id = bar.id

我希望能够将“foo”和“bar”字符串分解出来，以便我可以拥有类似的内容：

WITH
subq(my_arg) AS
    SELECT *
    FROM my_table
    WHERE field = my_arg
SELECT *
FROM subq('foo')
INNER JOIN subq('bar')
ON subq('foo').id = subq('foo').id

因为，foos和bars实际上比这要大得多，而且其中只有两个，所以维护起来有点困难。

我知道使用WITH子句可能无法做到这一点，但是避免多次编写此子查询的最佳解决方案是什么？这实际上可能很简单，但我对SQL很新......

感谢您的帮助。

Answer 1

看起来这可能是你想要的：

SELECT *
FROM my_table foo
JOIN my_table bar ON foo.id = bar.id
JOIN my_table baz ON foo.id = baz.id
WHERE foo.field = 'foo'
AND bar.field = 'bar'
AND baz.field = 'baz'

如果WITH子句做了很多（并且不值得重复）：

WITH cte AS SELECT * FROM mytable <with some complex SQL>
SELECT *
FROM cte foo
JOIN cte bar ON foo.id = bar.id
JOIN cte baz ON foo.id = baz.id
WHERE foo.field = 'foo'
AND bar.field = 'bar'
AND baz.field = 'baz'

Answer 2

试试这个：

WITH subq AS (
    SELECT *
    FROM my_table
)
SELECT *
  FROM subq s1
     , subq s2
 WHERE s1.id = s2.id
   AND s1.field = 'foo'
   AND s2.field = 'bar'

或者您可以使用pipelined这样的功能：

CREATE TYPE t_tf_tab IS TABLE OF MY_TABLE%ROWTYPE;

CREATE OR REPLACE FUNCTION get_table_vals (
    p_val IN VARCHAR
)
RETURN t_tf_tab 
PIPELINED 
AS
BEGIN
  FOR i IN (SELECT * FROM MY_TABLE WHERE FIELD = p_val)
    PIPE ROW(t_tf_row(i.id, i.field, ...));   
  END LOOP;
  RETURN;
END;

SELECT *
  FROM TABLE(get_table_vals('foo')) s1
     , TABLE(get_table_vals('bar')) s2
 where s1.id = s2.id

Answer 3

您可以在下一个表达式中重用WITH表达式。但据我所知，你不能参数化它。所以这可能会有所帮助：

WITH
foosAndbars AS 
    (SELECT *
     FROM [Very complex query] ).
foos AS (
    SELECT *
    FROM foosAndbars 
    WHERE field = 'foo'),
bars AS (
    SELECT *
    FROM foosAndbars 
    WHERE field = 'bar')
SELECT *
FROM foo
INNER JOIN bar
ON foo.id = bar.id

SQL：包含参数的WITH子句？

3 个答案: