PL Sql Creating Hierarcical ID - sql

Question

i have this table without any hierarcy but its obvious it has a hierarcy like 100,100-01,100-01-01 ..
ID ACCOUNT
3 100
4 100-01
5 100-01-01
6 100-01-01-001
7 100-01-01-006
8 100-01-01-271
9 100-01-01-272
10 100-01-01-273
11 100-01-16
12 100-01-16-420
13 100-02
14 100-02-01
15 100-02-01-001
16 100-02-01-271
17 100-02-01-272
i want to fill like this; 100-01 is child of 100 , 100-01-01 is child of 100-01 so ;
ID H_ID ACCOUNT
3 100
4 3 100-01
5 4 100-01-01
6 5 100-01-01-001
7 5 100-01-01-006
8 5 100-01-01-271
9 5 100-01-01-272
10 5 100-01-01-273
11 4 100-01-16
12 11 100-01-16-420
13 3 100-02
14 13 100-02-01
15 14 100-02-01-001
16 14 100-02-01-271
17 14 100-02-01-272
Thank you for your time...

Answer 1

You can find the 'parent' account with substr and instr:
select id, account, substr(account, 1, instr(account, '-', -1, 1) - 1)
from your_table;
ID ACCOUNT SUBSTR(ACCOUN
---------- ------------- -------------
3 100
4 100-01 100
5 100-01-01 100-01
6 100-01-01-001 100-01-01
...
You can then (outer) join the table to itself using that value:
select t1.id, t2.id as h_id, t1.account
from your_table t1
left join your_table t2
on t2.account = substr(t1.account, 1, instr(t1.account, '-', -1, 1) - 1)
order by t1.id;
which with your sample data gives:
ID H_ID ACCOUNT
---------- ---------- -------------
3 100
4 3 100-01
5 4 100-01-01
6 5 100-01-01-001
7 5 100-01-01-006
8 5 100-01-01-271
9 5 100-01-01-272
10 5 100-01-01-273
11 4 100-01-16
12 11 100-01-16-420
13 3 100-02
14 13 100-02-01
15 14 100-02-01-001
16 14 100-02-01-271
17 14 100-02-01-272
There is no parent for ID 3 so it isn't clear why your expected result has 2 for that.
You could also extract the 'branch' and 'leaf' parts of the account if that's useful:
select t1.id, t2.id as h_id, t1.account,
substr(t1.account, 1, instr(t1.account, '-', -1, 1) - 1) as branch,
substr(t1.account, instr(t1.account, '-', -1, 1) + 1) as leaf
from your_table t1
left join your_table t2
on t2.account = substr(t1.account, 1, instr(t1.account, '-', -1, 1) - 1)
order by t1.id;
ID H_ID ACCOUNT BRANCH LEAF
---------- ---------- ------------- ------------- -------------
3 100 100
4 3 100-01 100 01
5 4 100-01-01 100-01 01
6 5 100-01-01-001 100-01-01 001
7 5 100-01-01-006 100-01-01 006
8 5 100-01-01-271 100-01-01 271
...
db<>fiddle
In principle, if this is a real hierarchy, you only need to store the ID, parent ID and leaf part, and can then reconstruct the whole account with a hierarchical query (or recursive CTE). That may not actually be practical, or sensible, for your situation.
Looking at the edit history of your question, it seems like your table has an empty h_id column that you want to populate. You can use the same mechanism with a correlated update:
update your_table t1
set h_id = (
select t2.id
from your_table t2
where t2.account = substr(t1.account, 1, instr(t1.account, '-', -1, 1) - 1)
);
db<>fiddle