Extract the maximum value from the last number in a hierarchyID? - sql

Question

I have a column with hierarchy IDs converted to strings in SQL Server. I need to add new hierarcyIDs for the new lines, but first I have to find the last child of the current ID. The hierarchyIDs are look like these:
/1/1/1/6/1/
/1/1/1/6/7/
/1/1/1/6/3/
/1/1/1/6/13/
/1/1/1/6/4/
As you can see, the maximum number is not equal with the count of the lines, so I can not use count()+1 unfortunately.
What I need to extract from this list is:
13
I only have experience in PL SQL, where it was easy to do this with regexp functions, but I can not find the solution in SQL Server.

Answer 1

You can use some STRING operation as below to get your desired output-
DEMO HERE
WITH your_table(your_column)
AS
(
SELECT '/1/1/1/6/1/' UNION ALL
SELECT '/1/1/1/6/7/' UNION ALL
SELECT '/1/1/1/6/3/' UNION ALL
SELECT '/1/1/1/6/13/' UNION ALL
SELECT '/1/1/1/6/4/'
)
SELECT
MAX(
CAST(
REVERSE(
LEFT(
REVERSE(LEFT(your_column,LEN(your_column)-1)),
CHARINDEX('/',REVERSE(LEFT(your_column,LEN(your_column)-1)) ,0) - 1
)
)
AS INT
)
)
FROM your_table
Note: Data has to be as your sample data

Answer 2

The creators of hierarchyid have anticipated your needs and have an officially supported solution for this.
declare #h table (h HIERARCHYID);
insert into #h (h)
values
('/1/1/1/6/1/'),
('/1/1/1/6/7/'),
('/1/1/1/6/3/'),
('/1/1/1/6/13/'),
('/1/1/1/6/4/');
declare #parent HIERARCHYID = '/1/1/1/6/';
declare #maxChild HIERARCHYID = (
select max(h)
from #h
where h.IsDescendantOf(#parent) = 1
);
-- ToString() added here for readability in the output;
-- it's not needed to be used as data.
select #parent.GetDescendant(#maxChild, null).ToString();
You can read more about this here.
Another way around this is to specify your own components to the hierarchyid yourself. I like to use the primary key values. For example, let's say that this data represents a company's org chart. If EmployeeID 1 is the CEO, 42 is the CFO (who reports to the CEO) and 306 is Accounting Manager (who reports to the CFO), the latter's hierarchyid would be /1/42/306/. Because the PK values are unique, the generated hierarchid is also unique.

Answer 3

To give some ideas, 2 solutions.
The first one is for Sql Server 2017 and beyond.
The second for most versions below that.
create table YourTable
(
ID int identity(101,1) primary key,
hierarcyIDs varchar(100)
)
GO
✓
insert into YourTable
(hierarcyIDs) values
('/1/2/3/4/5/')
,('/1/2/3/4/15/')
,('/1/2/3/4/10/')
,('/11/12/13/14/15/')
,('/11/12/13/14/42/')
;
GO
5 rows affected
SELECT
[1] as id1,
[2] as id2,
[3] as id3,
[4] as id4,
max([5]) as id5,
concat_ws('/','',[1],[2],[3],[4],max([5])+1,'') as nextHierarcyIDs
FROM YourTable
OUTER APPLY
(
select *
from
( select try_cast(value as int) as id
, row_number() over (order by (select 0)) rn
from string_split(trim('/' from hierarcyIDs),'/') s
) src
pivot (max(id)
for rn in ([1],[2],[3],[4],[5])
) pvt
) anarchy
group by [1],[2],[3],[4]
GO
id1 | id2 | id3 | id4 | id5 | nextHierarcyIDs
--: | --: | --: | --: | --: | :---------------
1 | 2 | 3 | 4 | 15 | /1/2/3/4/16/
11 | 12 | 13 | 14 | 42 | /11/12/13/14/43/
SELECT hierarcyIdLvl1to4
, MAX(hierarcyIDLvl5) AS maxHierarcyIDLvl5
FROM
(
SELECT id, hierarcyIDs
, substring(hierarcyIDs, 0, len(hierarcyIDs)-charindex('/',
reverse(hierarcyIDs),2)+2) AS hierarcyIdLvl1to4
, reverse(substring(reverse(hierarcyIDs),2,charindex('/',
reverse(hierarcyIDs),2)-2)) AS hierarcyIDLvl5
FROM YourTable
) q
GROUP BY hierarcyIdLvl1to4
GO
hierarcyIdLvl1to4 | maxHierarcyIDLvl5
:---------------- | :----------------
/1/2/3/4/ | 5
/11/12/13/14/ | 42
db<>fiddle here