Find relation level from text field - sql

Question

I have a table containing geographical structure of units. There are parent-child relation columns but I want to use the existing text field (instead of recursion) to find the relation level between the items.
(here's a table creation script)
drop table if exists #temp_structure
create table #temp_structure
(org_id int,
parent_org_id int,
org_name nvarchar(255),
search_tree nvarchar(255))
insert into #temp_structure
values
(1,null,'World','| 1 |'),
(2,1,'Europe','| 1 | 2 |'),
(3,1,'North America','| 1 | 3 |'),
(4,1,'South America','| 1 | 4 |'),
(5,1,'Asia','| 1 | 5 |'),
(6,1,'Africa','| 1 | 6 |'),
(7,1,'Australia','| 1 | 7 |'),
(8,2,'Spain','| 1 | 2 | 8 |'),
(9,2,'Germany','| 1 | 2 | 9 |'),
(10,2,'Italy','| 1 | 2 | 10 |'),
(11,2,'France','| 1 | 2 | 11 |'),
(12,8,'Madrid ','| 1 | 2 | 8 | 12 |'),
(13,8,'Barcelona ','| 1 | 2 | 8 | 13 |'),
(14,9,'Berlin','| 1 | 2 | 9 | 14 |'),
(15,9,'Munich','| 1 | 2 | 9 | 15 |'),
(16,10,'Rome','| 1 | 2 | 10 | 16 |'),
(17,10,'Milano','| 1 | 2 | 10 | 17 |'),
(18,11,'Paris','| 1 | 2 | 11 | 18 |'),
(19,11,'Marseille','| 1 | 2 | 11 | 19 |')
The expected result I would like to achieve is presented below (I listed only one 4th level example):
+--------+-------------+------------+
| org_id | search_item | nest_level |
+--------+-------------+------------+
| 1 | 1 | 1 |
| 2 | 2 | 1 |
| 2 | 1 | 2 |
| 3 | 3 | 1 |
| 3 | 1 | 2 |
| 4 | 4 | 1 |
| 4 | 1 | 2 |
| 5 | 5 | 1 |
| 5 | 1 | 2 |
| 6 | 6 | 1 |
| 6 | 1 | 2 |
| 7 | 7 | 1 |
| 7 | 1 | 2 |
| 8 | 8 | 1 |
| 8 | 2 | 2 |
| 8 | 1 | 3 |
| 9 | 9 | 1 |
| 9 | 2 | 2 |
| 9 | 1 | 3 |
| 10 | 10 | 1 |
| 10 | 2 | 2 |
| 10 | 1 | 3 |
| 11 | 11 | 1 |
| 11 | 2 | 2 |
| 11 | 1 | 3 |
| 12 | 12 | 1 |
| 12 | 8 | 2 |
| 12 | 2 | 3 |
| 12 | 1 | 4 |
.....................................
+--------+-------------+------------+
I was able to pull the org_id-search_item relation using STRING_SPLIT, but I still miss the tricky level part (I wonder about enumerating the '|' characters)
SELECT t.org_id
--,substring(replace(search_tree, ' ', ''), 2, len(replace(search_tree, ' ', '')) - 2)
,ss.value as search_item
FROM #temp_structure t
CROSS APPLY string_split(substring(replace(search_tree, ' ', ''), 2, len(replace(search_tree, ' ', '')) - 2),'|') ss

Answer 1

I have not thoroughly tested this, but you could try something like the following:
-- Table mock-up.
DECLARE #temp TABLE ( org_id int, parent_org_id int, org_name nvarchar(255), search_tree nvarchar(255) )
-- Insert sample data...
INSERT INTO #temp VALUES
(1,null,'World','| 1 |'),(2,1,'Europe','| 1 | 2 |'),
(3,1,'North America','| 1 | 3 |'),(4,1,'South America','| 1 | 4 |'),
(5,1,'Asia','| 1 | 5 |'),(6,1,'Africa','| 1 | 6 |'),
(7,1,'Australia','| 1 | 7 |'),(8,2,'Spain','| 1 | 2 | 8 |'),
(9,2,'Germany','| 1 | 2 | 9 |'),(10,2,'Italy','| 1 | 2 | 10 |'),
(11,2,'France','| 1 | 2 | 11 |'),(12,8,'Madrid ','| 1 | 2 | 8 | 12 |');
-- Select data in a nested level...
SELECT
org_id,
search_item,
ROW_NUMBER() OVER ( PARTITION BY org_id ORDER BY org_id, parent_org_id, search_item DESC ) AS nest_level
FROM #temp AS tmp
CROSS APPLY (
SELECT CAST ( [value] AS INT ) AS search_item FROM STRING_SPLIT ( tmp.search_tree, '|' )
WHERE NULLIF ( [value], '' ) IS NOT NULL
) AS tree
ORDER BY
org_id, parent_org_id, search_item DESC;
Returns
+--------+-------------+------------+
| org_id | search_item | nest_level |
+--------+-------------+------------+
| 1 | 1 | 1 |
| 2 | 2 | 1 |
| 2 | 1 | 2 |
| 3 | 3 | 1 |
| 3 | 1 | 2 |
| 4 | 4 | 1 |
| 4 | 1 | 2 |
| 5 | 5 | 1 |
| 5 | 1 | 2 |
| 6 | 6 | 1 |
| 6 | 1 | 2 |
| 7 | 7 | 1 |
| 7 | 1 | 2 |
| 8 | 8 | 1 |
| 8 | 2 | 2 |
| 8 | 1 | 3 |
| 9 | 9 | 1 |
| 9 | 2 | 2 |
| 9 | 1 | 3 |
| 10 | 10 | 1 |
| 10 | 2 | 2 |
| 10 | 1 | 3 |
| 11 | 11 | 1 |
| 11 | 2 | 2 |
| 11 | 1 | 3 |
| 12 | 12 | 1 |
| 12 | 8 | 2 |
| 12 | 2 | 3 |
| 12 | 1 | 4 |
+--------+-------------+------------+