问题
I've created some sort of "dependencies" table that contain the entire dependencies of ours night process.
The table looks like this:
GRAND_MODEL | WAIT_4_MODEL_NAME
test test1
test test2
test test3
test2 test3
test3 test4
test4 test5
This table means -> test needs to wait for test1 , test2, test3 in order to finish, but also it needs to wait for test4 and test5 , because test3/4 waits for them .
test1 doesn't wait for anything , test2 waits for test3 and therefore also for test4 and therefore test5 .
So the result should look something like this:
FIRST_MODEL | SECOND_MODEL | THIRD_MODEL | FORTH_MODEL | FIFTH_MODEL | SIXTH_MODEL
test5 test4 test3 test2 test NULL
test5 test4 test3 test NULL NULL
test4 test3 test2 test NULL NULL
.................................
What I've tried :
SELECT distinct prior wait_4_model_name as first_m,
wait_4_model_name as second_m,
grand_model as third_m
from (SELECT distinct grand_model, wait_4_model_name
FROM DEL_SAGI_FOR_HIERARCHY)
connect by NOCYCLE prior grand_model = wait_4_model_name
But that only generates the first level of the hierarchy.
Thanks in advance.
EDIT: Note that there can be opposite dependencies , test waits for test1 and test1 waits for test (each model is big, so it's possible for a part of the model to wait for a part of another model)
回答1:
Oracle Setup:
CREATE TABLE table_name ( GRAND_MODEL, WAIT_4_MODEL_NAME ) AS
SELECT 'test', 'test1' FROM DUAL UNION ALL
SELECT 'test', 'test2' FROM DUAL UNION ALL
SELECT 'test', 'test3' FROM DUAL UNION ALL
SELECT 'test2', 'test3' FROM DUAL UNION ALL
SELECT 'test3', 'test4' FROM DUAL UNION ALL
SELECT 'test4', 'test5' FROM DUAL;
Query:
SELECT REGEXP_SUBSTR( tests, '[^|]+', 1, 1 ) AS first_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 2 ) AS second_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 3 ) AS third_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 4 ) AS fourth_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 5 ) AS fifth_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 6 ) AS sixth_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 7 ) AS seventh_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 8 ) AS eighth_model,
REGEXP_SUBSTR( tests, '[^|]+', 1, 9 ) AS ninth_model
FROM (
SELECT SYS_CONNECT_BY_PATH( wait_4_model_name, '|' ) || '|' || grand_model AS tests
FROM table_name
CONNECT BY PRIOR grand_model = wait_4_model_name
);
Output:
FIRST_MODEL SECOND_MODEL THIRD_MODEL FOURTH_MODEL FIFTH_MODEL SIXTH_MODEL SEVENTH_MODEL EIGHTH_MODEL NINTH_MODEL
----------- ------------ ----------- ------------ ----------- ----------- ------------- ------------ -----------
test1 test
test2 test
test3 test
test3 test2
test3 test2 test
test4 test3
test4 test3 test
test4 test3 test2
test4 test3 test2 test
test5 test4
test5 test4 test3
test5 test4 test3 test
test5 test4 test3 test2
test5 test4 test3 test2 test
回答2:
Here is a solution that does not require string concatenation and re-split:
select a.wait_4_model_name model1,
a.grand_model model2,
b.grand_model model3,
c.grand_model model4,
d.grand_model model5,
e.grand_model model6
from DEL_SAGI_FOR_HIERARCHY a
left join DEL_SAGI_FOR_HIERARCHY b ON a.grand_model = b.wait_4_model_name
left join DEL_SAGI_FOR_HIERARCHY c ON b.grand_model = c.wait_4_model_name
left join DEL_SAGI_FOR_HIERARCHY d ON c.grand_model = d.wait_4_model_name
left join DEL_SAGI_FOR_HIERARCHY e ON d.grand_model = e.wait_4_model_name
where a.wait_4_model_name not in (
select grand_model from DEL_SAGI_FOR_HIERARCHY)
order by 1, 2, 3, 4, 5, 6
Output for the sample data is:
+--------+--------+--------+--------+--------+--------+
| MODEL1 | MODEL2 | MODEL3 | MODEL4 | MODEL5 | MODEL6 |
+--------+--------+--------+--------+--------+--------+
| test1 | test | - | - | - | - |
| test5 | test4 | test3 | test | - | - |
| test5 | test4 | test3 | test2 | test | - |
+--------+--------+--------+--------+--------+--------+
Note that your sample data has no instances of models that do not have any dependencies at all, i.e. that do not need to wait for another model, nor are necessary for any models before they can start.
来源：https://stackoverflow.com/questions/38550821/hierarchical-queries