SQL Server : GROUP CONCAT with DISTINCT is sorting natural data input - sql

Question

I have a similar situation. I start out with a table that has data input into a column from another source. This data is comma delimited coming in. I need to manipulate the data to remove a section at the end of each. So I split the data and remove the end with the code below. (I added the ID column later to be able to sort. I also added WITH SCHEMABINDING later to add an XML index but nothing works. I can remove this ... and the ID column, but I do not see any difference one way or the other):
ALTER VIEW [dbo].[vw_Routing]
WITH SCHEMABINDING
AS
SELECT TOP 99.9999 PERCENT
ROW_NUMBER() OVER (ORDER BY CableID) - 1 AS ID,
CableID AS [CableID],
SUBSTRING(m.n.value('.[1]', 'varchar(8000)'), 1, 13) AS Routing
FROM
(SELECT
CableID,
CAST('<XMLRoot><RowData>' + REPLACE([RouteNodeList], ',', '</RowData><RowData>') + '</RowData></XMLRoot>' AS xml) AS x
FROM
[dbo].[Cables]) t
CROSS APPLY
x.nodes('/XMLRoot/RowData') m (n)
ORDER BY
ID)
Now I need to concatenate data from the Routing column's rows into one row grouped by another column into a column again. I have the code working except that it is reordering my data; I must have the data in the order it is input into the table as it is Cable Routing information. I must also remove duplicates. I use the following code. The SELECT DISTINCT removes the duplicates, but reorders the data. The SELECT (without DISTINCT) keeps the correct data order, but does NOT remove the duplicates:
Substring(
(
SELECT DISTINCT ','+ x3.Routing AS [text()] --This DISTINCT reorders the routes once concatenated.
--SELECT ','+ x3.Routing AS [text()] --This without the DISTINCT does not remove duplicates.
From vw_Routing x3
Where x3.CableID = c.CableId
For XML PATH ('')
), 2, 1000) [Routing],
I tried the code you gave above and it provided the same results with the DISTINCT reordering the data but without DISTINCT not removing the duplicates.

Answer 1

Perhaps GROUP BY with ORDER BY will work:
stuff((select ','+ x3.Routing AS [text()] --This DISTINCT reorders the routes once concatenated.
--SELECT ','+ x3.Routing AS [text()] --This without the DISTINCT does not remove duplicates.
from vw_Routing x3
where x3.CableID = c.CableId
group by x3.Routing
order by min(x3.id)
for XML PATH ('')
), 1, 1, '') as [Routing],
I also replaced the SUBSTRING() with STUFF(). The latter is more standard for this operation.

Answer 2

To https://stackoverflow.com/users/1144035/gordon-linoff
Unfortunately, that did not work. It gave me the same result as my select statement; that is, no dups but reordering data.
HOWEVER, I found the correct answer earlier today:
I figured it out finally!! I still have to get implement it within the other code and add the new Cable Area code, but the hard part it over!!!!!
I am going to post the following to the forums so that they know not to work on it .... I was writing this to send to my friend for his help, but I figured it out myself before I sent it.
I started with raw, comma separated data in the records of a table … the data is from another source. I had to remove some of the information from each value, so I used the following code to split it up and manipulate it:
Code1
Once that was done, I had to put the manipulated data back in the same form in the same order and with no duplicates. So I needed a SELECT DISTINCT. When I used the commented out SELECT DISTINCT below, it removed duplicates but it changed the order of the data which I could not have as it is Cable Tray Routing Data. When I took out the SELECT DISTINCT, it kept correct order, but left duplicates.
Because I was using XML PATH, I had to change this code …… To this code so that I could use SELECT DISTINCT remove the duplicates:Code2 and Code3
