我试图创建一个Oracle SQL语句,以从不同级别的多个父级子记录中获取顶级根级父记录。表格结构如下。下面的顶级根父母是parent_membership_id 53887,这个父记录有许多孩子,他们也是其他孩子的父母。我想要的是一个查询,如果我查询说200326的成员,查询带回根成员53887,或者如果我查询200322我得到根成员53887.我想你知道我想要做什么。谢谢你。
CREATE TABLE MEMBERSHIP_LINK
( MEMBERSHIP_LINK_ID NUMBER(10) NOT NULL,
CHILD_MEMBERSHIP_ID NUMBER(10) NOT NULL,
PARENT_MEMBERSHIP_ID NUMBER(10) NOT NULL);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(35, 53890, 53887);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(24475, 200322, 53887);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(24476, 200322, 53887);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(34, 53889, 53888);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(5941, 112177, 53889);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(33, 53888, 53890);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(24477, 200323, 200322);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(24478, 200323, 200322);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(24479, 200325, 200323);
Insert into MEMBERSHIP_LINK
(MEMBERSHIP_LINK_ID, CHILD_MEMBERSHIP_ID, PARENT_MEMBERSHIP_ID)
Values
(24480, 200326, 200323);
COMMIT;
答案 0 :(得分:0)
解决了问题。 SQL在下面。
SELECT DISTINCT meli.parent_membership_id
FROM MEMBERSHIP_LINK meli
WHERE LEVEL = ( SELECT max(level)
FROM MEMBERSHIP_LINK meli_in
START WITH meli_in.child_membership_id = :membership_id
CONNECT BY meli_in.child_membership_id = PRIOR meli_in.parent_membership_id )
START WITH meli.child_membership_id = :membership_id
CONNECT BY meli.child_membership_id = prior meli.parent_membership_id
答案 1 :(得分:0)
这个有效,但由于记录24477和24478定义了相同的关系,它需要不同的运算符。
SELECT DISTINCT ML.PARENT_MEMBERSHIP_ID
FROM MEMBERSHIP_LINK ml
WHERE CONNECT_BY_ISLEAF = 1 -- Limit to the "Root" element(s)
START WITH ML.CHILD_MEMBERSHIP_ID = :decendent_id
CONNECT BY ML.CHILD_MEMBERSHIP_ID = prior ML.PARENT_MEMBERSHIP_ID;
将PRIOR
关键字放在CONNECT BY
上的PARENT_MEMBERSHIP_ID
子句中会导致树遍历到根,并使" root"节点离开。
如果您正在尝试寻找共享祖先,则需要采用不同的方法:
WITH ancestry AS
( SELECT DISTINCT CONNECT_BY_ROOT CHILD_MEMBERSHIP_ID child_id
, ML.PARENT_MEMBERSHIP_ID ancestor_id
, level generation_gap
, CONNECT_BY_ISLEAF
FROM MEMBERSHIP_LINK ml
START WITH ML.CHILD_MEMBERSHIP_ID in (:Descendent_ID1,:Descendent_ID2)
CONNECT BY ML.CHILD_MEMBERSHIP_ID = prior ML.PARENT_MEMBERSHIP_ID
)
SELECT ancestor_id
FROM ancestry
WHERE child_id = :Descendent_ID1
INTERSECT
SELECT ancestor_id
FROM ancestry
WHERE child_id = :Descendent_ID2;
从这里你可以确定最新的(最年轻的)共享祖先,最早的共享祖先和共同的血统:
WITH ancestry AS (
SELECT DISTINCT CONNECT_BY_ROOT CHILD_MEMBERSHIP_ID child_id
, ML.PARENT_MEMBERSHIP_ID ancestor_id
, level generation_gap
, CONNECT_BY_ISLEAF
FROM MEMBERSHIP_LINK ml
START WITH ML.CHILD_MEMBERSHIP_ID in (:Descendent_ID1,:Descendent_ID2)
CONNECT BY ML.CHILD_MEMBERSHIP_ID = prior ML.PARENT_MEMBERSHIP_ID
), common AS (
SELECT ancestor_id
FROM ancestry
WHERE child_id = :Descendent_ID1
INTERSECT
SELECT ancestor_id
FROM ancestry
WHERE child_id = :Descendent_ID2
)
SELECT MIN( a.ANCESTOR_ID ) keep( dense_rank FIRST ORDER BY a.GENERATION_GAP ) Youngest_Ancestor
, LISTAGG(a.ANCESTOR_ID, '->') within group (order by a.GENERATION_GAP) common_lineage
, MIN( a.ANCESTOR_ID ) keep( dense_rank FIRST ORDER BY a.GENERATION_GAP desc ) Oldest_Ancestor
FROM ancestry a
JOIN common c
ON a.ancestor_id = c.ancestor_id
WHERE a.child_id = :Descendent_ID1;