SQL Trigger to split string during insert without a common delimiter and store it into another table - sql

Question

Currently I have a system that is dumping data into a table with the format:
Table1
Id#, row#, row_dump
222, 1, “set1 = aaaa set2 =aaaaaa aaaa dd set4=1111”
I want to take the row dump and transpose it into rows and insert it into another table of the format:
Table2
Id#, setting, value
222, ‘set1’,’aaa’
222, ‘set2’,’aaaaaa aaaa dd’
222, ‘set4’,’1111’
Is there a way to make a trigger in MSSQL that will parse this string on insert in Table1 and insert it into Table2 properly?
All of the examples I’ve found required a common delimiter. ‘=’ separates the setting from the value, space(s) separate a value from a setting but a value could have spaces in it (settings do not have spaces in them so the last word before the equal sign is the setting name but there could be spaces between the setting name and equal sign).
There could be 1-5 settings and values in any given row. The values can have spaces. There may or may not be space between the setting name and the ‘=’ sign.
I have no control over the original insert process or format as it is used for other purposes.

Answer 1

You could use 'set' as a delimiter. This is a simple sample. It obviously may have to be molded to your environment.
use tempdb
GO
IF OBJECT_ID('dbo.fn_TVF_Split') IS NOT NULL
DROP FUNCTION dbo.fn_TVF_Split;
GO
CREATE FUNCTION dbo.fn_TVF_Split(#arr AS NVARCHAR(2000), #sep AS NCHAR(3))
RETURNS TABLE
AS
RETURN
WITH
L0 AS (SELECT 1 AS C UNION ALL SELECT 1) --2 rows
,L1 AS (SELECT 1 AS C FROM L0 AS A, L0 AS B) --4 rows (2x2)
,L2 AS (SELECT 1 AS C FROM L1 AS A, L1 AS B) --16 rows (4x4)
,L3 AS (SELECT 1 AS C FROM L2 AS A, L2 AS B) --256 rows (16x16)
,L4 AS (SELECT 1 AS C FROM L3 AS A, L3 AS B) --65536 rows (256x256)
,L5 AS (SELECT 1 AS C FROM L4 AS A, L4 AS B) --4,294,967,296 rows (65536x65536)
,Nums AS (SELECT row_number() OVER (ORDER BY (SELECT 0)) AS N FROM L5)
SELECT
(n - 1) - LEN(REPLACE(LEFT(#arr, n-1), #sep, N'')) + 1 AS pos,
SUBSTRING(#arr, n, CHARINDEX(#sep, #arr + #sep, n) - n) AS element
FROM Nums
WHERE
n <= LEN(#arr) + 3
AND SUBSTRING(#sep + #arr, n, 3) = #sep
AND N<=100000
GO
declare #t table(
Id int,
row int,
row_dump varchar(Max)
);
insert into #t values(222, 1, 'set1 = aaaa set2 =aaaaaa aaaa dd set4=1111')
insert into #t values(111, 2, ' set1 =cx set2 =4444set4=124')
DECLARE #t2 TABLE(
Id int,
Setting VARCHAR(6),
[Value] VARCHAR(50)
)
insert into #t2 (Id,Setting,Value)
select
Id,
[Setting]='set' + left(LTRIM(element),1),
[Value]=RIGHT(element,charindex('=',reverse(element))-1)
from #t t
cross apply dbo.fn_TVF_Split(row_dump,'set')
where pos > 1
order by
id asc,
'set' + left(LTRIM(element),1) asc
select *
from #t2
Update
You could do something like this. It is not optimal and could probably be better handled in the transformation tool or application. Anyway here we go.
Note: You will need the split function I posted before.
declare #t table(
Id int,
row int,
row_dump varchar(Max)
);
insert into #t values(222, 1, 'set1 = aaaa set2 =aaaaaa aaaa dd set3=abc set4=1111 set5=7373')
insert into #t values(111, 2, 'set1 =cx set2 = 4444 set4=124')
DECLARE #t2 TABLE(
Id int,
Setting VARCHAR(6),
[Value] VARCHAR(50)
)
if OBJECT_ID('tempdb.dbo.#Vals') IS NOT NULL
BEGIN
DROP TABLE #Vals;
END
CREATE TABLE #Vals(
Id INT,
Row INT,
Element VARCHAR(MAX),
pos int,
value VARCHAR(MAX)
);
insert into #Vals
select
Id,
row,
element,
pos,
Value=STUFF(LEFT(element,len(element) - CHARINDEX(' ',reverse(element))),1,1,'')
from(
select
Id,
row,
row_dump = REPLACE(REPLACE(REPLACE(row_dump,'= ','='),' =','='),'=','=|')
from #t
) AS t
cross apply dbo.fn_TVF_Split(row_dump,'=')
where pos >=1 and pos < 10
insert into #t2 (Id,Setting,Value)
select
t1.Id,
Setting =
(
SELECT TOP 1
CASE WHEN t2.pos = 1
THEN LTRIM(RTRIM(t2.element))
ELSE LTRIM(RTRIM(RIGHT(t2.element,CHARINDEX(' ',REVERSE(t2.element)))))
END
FROM #Vals t2
where
t2.Id = t1.id
and t2.row = t1.row
and t2.pos < t1.pos
ORDER BY t2.pos DESC
),
t1.Value
from #Vals t1
where t1.pos > 1 and t1.pos < 10
order by t1.id,t1.row,t1.pos
select * from #t2