Question: I want to write a custom aggregate function that concatenates string on group by.
So that I can do a
SELECT SUM(FIELD1) as f1, MYCONCAT(FIELD2) as f2
FROM TABLE_XY
GROUP BY FIELD1, FIELD2
All I find is SQL CRL aggregate functions, but I need SQL, without CLR.
Edit:1
The query should look like this:
SELECT SUM(FIELD1) as f1, MYCONCAT(FIELD2) as f2
FROM TABLE_XY
GROUP BY FIELD0
Edit 2:
It is true that it isn't possible without CLR.
However, the subselect answer by astander can be modified so it doesn't XML-encode special characters.
The subtle change for this is to add this after "FOR XML PATH":
,
TYPE
).value('.[1]', 'nvarchar(MAX)')
Here a few examples
DECLARE #tT table([A] varchar(200), [B] varchar(200));
INSERT INTO #tT VALUES ('T_A', 'C_A');
INSERT INTO #tT VALUES ('T_A', 'C_B');
INSERT INTO #tT VALUES ('T_B', 'C_A');
INSERT INTO #tT VALUES ('T_C', 'C_A');
INSERT INTO #tT VALUES ('T_C', 'C_B');
INSERT INTO #tT VALUES ('T_C', 'C_C');
SELECT
A AS [A]
,
(
STUFF
(
(
SELECT DISTINCT
', ' + tempT.B AS wtf
FROM #tT AS tempT
WHERE (1=1)
--AND tempT.TT_Status = 1
AND tempT.A = myT.A
ORDER BY wtf
FOR XML PATH, TYPE
).value('.[1]', 'nvarchar(MAX)')
, 1, 2, ''
)
) AS [B]
FROM #tT AS myT
GROUP BY A
SELECT
(
SELECT
',äöü<>' + RM_NR AS [text()]
FROM T_Room
WHERE RM_Status = 1
ORDER BY RM_NR
FOR XML PATH('')
) AS XmlEncodedNoNothing
,
SUBSTRING
(
(
SELECT
',äöü<>' + RM_NR AS [data()]
FROM T_Room
WHERE RM_Status = 1
ORDER BY RM_NR
FOR XML PATH('')
)
,2
,10000
) AS XmlEncodedSubstring
,
(
STUFF
(
(
SELECT ',äöü<>' + RM_NR + CHAR(10)
FROM T_Room
WHERE RM_Status = 1
ORDER BY RM_NR
FOR XML PATH, TYPE
).value('.[1]', 'nvarchar(MAX)')
, 1, 1, ''
)
) AS XmlDecodedStuffInsteadSubstring
You cannot write custom aggregates outside of the CLR.
The only type of functions you can write in pure T-SQL are scalar and table valued functions.
Compare the pages for CREATE AGGREGATE, which only lists CLR style options, with CREATE FUNCTION, which shows T-SQL and CLR options.
Have a look at something like. This is not an aggregate function. If you wish to implement your own aggregate function, it will have to be CLR...
DECLARE #Table TABLE(
ID INT,
Val VARCHAR(50)
)
INSERT INTO #Table (ID,Val) SELECT 1, 'A'
INSERT INTO #Table (ID,Val) SELECT 1, 'B'
INSERT INTO #Table (ID,Val) SELECT 1, 'C'
INSERT INTO #Table (ID,Val) SELECT 2, 'B'
INSERT INTO #Table (ID,Val) SELECT 2, 'C'
--Concat
SELECT t.ID,
SUM(t.ID),
stuff(
(
select ',' + t1.Val
from #Table t1
where t1.ID = t.ID
order by t1.Val
for xml path('')
),1,1,'') Concats
FROM #Table t
GROUP BY t.ID
Starting from 2017 there is built-in concatenate aggregate function STRING_AGG :)
https://learn.microsoft.com/en-us/sql/t-sql/functions/string-agg-transact-sql?view=sql-server-2017
Found this link around concatenation which covers methods like
Concatenating values when the number of items are not known
Recursive CTE method
The blackbox XML methods
Using Common Language Runtime
Scalar UDF with recursion
Table valued UDF with a WHILE loop
Dynamic SQL
The Cursor approach
Non-reliable approaches
Scalar UDF with t-SQL update extension
Scalar UDF with variable concatenation in SELECT
Though it doesn't cover aggerate functions there may be some use around concatenation in there to help you with your problem.
This solution works with no need of deploy from Visual studio or dll file in server.
Copy-Paste and it Work!
https://github.com/orlando-colamatteo/ms-sql-server-group-concat-sqlclr
dbo.GROUP_CONCAT(VALUE )
dbo.GROUP_CONCAT_D(VALUE ), DELIMITER )
dbo.GROUP_CONCAT_DS(VALUE , DELIMITER , SORT_ORDER )
dbo.GROUP_CONCAT_S(VALUE , SORT_ORDER )
You could do something like what I have done below to create a custom aggregate concatenation function in pure T-SQL. Obviously I have gone with a hard coded table name and group by column but it should illustrate the approach. There is probably some way to make this a truly generic function using dynamic TSQL constructed from input parameters.
/*
User defined function to help perform concatenations as an aggregate function
Based on AdventureWorks2008R2 SalesOrderDetail table
*/
--select * from sales.SalesOrderDetail
IF EXISTS (SELECT *
FROM sysobjects
WHERE name = N'fnConcatenate')
DROP FUNCTION fnConcatenate
GO
CREATE FUNCTION fnConcatenate
(
#GroupByValue int
)
returnS varchar(8000)
as
BEGIN
DECLARE #SqlString varchar(8000)
Declare #TempStore varchar(25)
select #SqlString =''
Declare #MyCursor as Cursor
SET #MyCursor = CURSOR FAST_FORWARD
FOR
Select ProductID
From sales.SalesOrderDetail where SalesOrderID = #GroupByValue
order by SalesOrderDetailID asc
OPEN #MyCursor
FETCH NEXT FROM #MyCursor
INTO #TempStore
WHILE ##FETCH_STATUS = 0
BEGIN
select #SqlString = ltrim(rtrim(#TempStore )) +',' + ltrim(rtrim(#SqlString))
FETCH NEXT FROM #MyCursor INTO #TempStore
END
CLOSE #MyCursor
DEALLOCATE #MyCursor
RETURN #SqlString
END
GO
select SalesOrderID, Sum(OrderQty), COUNT(*) as DetailCount , dbo.fnConcatenate(salesOrderID) as ConCatenatedProductList
from sales.SalesOrderDetail
where salesOrderID= 56805
group by SalesOrderID
Related
I have a query that returns a row
SELECT *
FROM table
WHERE id = 1;
I want to save the result into a nvarchar sql variable. I have seen similar questions Convert SQL Server result set into string but they only use select with the name of the columns, never with *.
select *
from table
where id = 1
for xml path ('')
However the answer is <column1>value1</column1> <column2>value2</column2> and I just want it to be value1, value2
Is there a way to achieve this? thank you!
If open to a helper function.
This will convert virtually any row, table or query to a string (delimited or not).
In the following examples I selected a PIPE delimiter with a CRLF line terminator.
Please note the usage and placement of _RN when a line terminator is required. Also note the ,ELEMENTS XSINIL ... this will included null values as empty string. If you want to exclude null values, simply omit the ,ELEMENTS XSINIL
Example as Entire Table or dbFiddle
Declare #YourTable Table (id int,[col_1] varchar(50),[col_2] varchar(50),[col_3] varchar(50),[col_n] varchar(50)) Insert Into #YourTable Values
(1,'data1','data2','data3','data4')
,(2,'data5','data6','data7','data8')
-- Entire Table
Declare #XML xml = (Select *,_RN=Row_Number() over (Order By (Select null)) From #YourTable for XML RAW,ELEMENTS XSINIL )
Select [dbo].[svf-str-Data-To-Delimited]('|',char(13)+char(10),#XML)
Returns
1|data1|data2|data3|data4
2|data5|data6|data7|data8
Example as Row Based
Select A.ID
,AsAString = [dbo].[svf-str-Data-To-Delimited]('|',char(13)+char(10),B.XMLData)
From #YourTable A
Cross Apply ( values ( (select a.* for xml RAW,ELEMENTS XSINIL )) )B(XMLData)
Returns
ID AsAString
1 1|data1|data2|data3|data4
2 2|data5|data6|data7|data8
The Function if Interested
CREATE Function [dbo].[svf-str-Data-To-Delimited] (#Delim varchar(50),#EOL varchar(50),#XML xml)
Returns varchar(max)
Begin
Return(
Select convert(nvarchar(max),(
Select case when Item='_RN' then ''
else case when nullif(lead(Item,1) over (Order by Seq),'_RN') is not null
then concat(Value,#Delim)
else concat(Value,#EOL)
end
end
From (
Select Seq = row_number() over(order by (select null))
,Item = xAttr.value('local-name(.)', 'nvarchar(100)')
,Value = xAttr.value('.','nvarchar(max)')
From #XML.nodes('/row/*') xNode(xAttr)
) A
Order By Seq
For XML Path (''),TYPE).value('.', 'nvarchar(max)') )
)
End
You can easily store the result as an XML string:
select *
from (values (1, 'x', getdate())) v(id, a, b)
where id = 1
for xml path ('');
Or as a JSON string:
select *
from (values (1, 'x', getdate())) v(id, a, b)
where id = 1
for json auto;
If you don't mind Using dynamic SQL (and INFORMATION_SCHEMA dictionary), for example, for SQL Server this works:
DECLARE #sql nvarchar(max) = '',
#result nvarchar(max),
#id int = 1
SELECT #sql += '+'',''+convert(nvarchar,' + QUOTENAME(column_name) +')' from INFORMATION_SCHEMA.columns where table_name = 'Student'
SET #sql = 'select #result=' + stuff(#sql,1,5,'') + ' from student where id = ' + CAST(#id as nvarchar)
EXECUTE sp_executesql #sql, N'#result nvarchar(max) OUTPUT', #result=#result OUTPUT
SELECT #result as MyOutput
I have following 2 tables in SQL Server
Category table:
Category
--------------------------
Delivery;Gauges;Book;Table
Category id:
id name
-----------------
13183 Delivery
88781 Gauges
88782 Book
12512 Table
Intended result is to have category table replaced with category id, as:
Category
-----------------------
13183;88781;88782;12512
I approached this by first separating category columns into separate columns using :
ltrim(rtrim(xDim.value('/x[1]','varchar(max)')))
ltrim(rtrim(xDim.value('/x[2]','varchar(max)')))
and so on. Then used left join and replace on each new column. Isn't there an easier way to do this? I searched on the net and stackoverflow but can't seem to find anything similar.
You can try to make a function to split your string value by a character.
CREATE FUNCTION Split_fun
( #Words nvarchar(MAX)
, #splitStr varchar(50)
)
RETURNS #Result_Table TABLE
(
[word] nvarchar(max) NULL
)
BEGIN
Declare #TempStr nvarchar(MAX)
WHILE (CHARINDEX(#splitStr,#Words)>0)
BEGIN
Set #TempStr=SUBSTRING(#Words,1,CHARINDEX(#splitStr,#Words)-1)
Insert into #Result_Table (word) Values (#TempStr)
Set #Words = REPLACE(#Words,#TempStr+#splitStr,'')
END/*End While*/
IF(LEN(RTRIM(LTRIM(#Words)))>0 And CHARINDEX(#splitStr,RTRIM(LTRIM(#Words)))=0)
Begin
Set #TempStr=#Words
Insert into #Result_Table (word) Values (#TempStr)
End
RETURN
END
you can use this function to make a result set by ';'.
do self-join with Category id table.
final you can use FOR XML connect all string by ; to get your expectation result.
;with cte as (
SELECT id
FROM T CROSS APPLY Split_fun(Category,';') v
JOIN T1 on v.word = t1.Category
)
select STUFF((
select distinct ';'+ cast(id as varchar(10))
FROM cte
FOR XML PATH(''), TYPE).value('.', 'NVARCHAR(MAX)'), 1, 1, '')
sqlfiddle
I have a field which is a concatenation of single letters. I am trying to order these strings within a view. These values can't be hard coded as there are too many. Is someone able to provide some guidance on the function to use to achieve the desired output below? I am using MSSQL.
Current output
CustID | Code
123 | BCA
Desired output
CustID | Code
123 | ABC
I have tried using a UDF
CREATE FUNCTION [dbo].[Alphaorder] (#str VARCHAR(50))
returns VARCHAR(50)
BEGIN
DECLARE #len INT,
#cnt INT =1,
#str1 VARCHAR(50)='',
#output VARCHAR(50)=''
SELECT #len = Len(#str)
WHILE #cnt <= #len
BEGIN
SELECT #str1 += Substring(#str, #cnt, 1) + ','
SET #cnt+=1
END
SELECT #str1 = LEFT(#str1, Len(#str1) - 1)
SELECT #output += Sp_data
FROM (SELECT Split.a.value('.', 'VARCHAR(100)') Sp_data
FROM (SELECT Cast ('<M>' + Replace(#str1, ',', '</M><M>') + '</M>' AS XML) AS Data) AS A
CROSS APPLY Data.nodes ('/M') AS Split(a)) A
ORDER BY Sp_data
RETURN #output
END
This works when calling one field
ie.
Select CustID, dbo.alphaorder(Code)
from dbo.source
where custid = 123
however when i try to apply this to top(10) i receive the error
"Invalid length parameter passed to the LEFT or SUBSTRING function."
Keeping in mind my source has ~4million records, is this still the best solution?
Unfortunately i am not able to normalize the data into a separate table with records for each Code.
This doesn't rely on a id column to join with itself, performance is almost as fast
as the answer by #Shnugo:
SELECT
CustID,
(
SELECT
chr
FROM
(SELECT TOP(LEN(Code))
SUBSTRING(Code,ROW_NUMBER() OVER(ORDER BY (SELECT NULL)),1)
FROM sys.messages) A(Chr)
ORDER by chr
FOR XML PATH(''), type).value('.', 'varchar(max)'
) As CODE
FROM
source t
First of all: Avoid loops...
You can try this:
DECLARE #tbl TABLE(ID INT IDENTITY, YourString VARCHAR(100));
INSERT INTO #tbl VALUES ('ABC')
,('JSKEzXO')
,('QKEvYUJMKRC');
--the cte will create a list of all your strings separated in single characters.
--You can check the output with a simple SELECT * FROM SeparatedCharacters instead of the actual SELECT
WITH SeparatedCharacters AS
(
SELECT *
FROM #tbl
CROSS APPLY
(SELECT TOP(LEN(YourString)) ROW_NUMBER() OVER(ORDER BY (SELECT NULL)) FROM master..spt_values) A(Nmbr)
CROSS APPLY
(SELECT SUBSTRING(YourString,Nmbr,1))B(Chr)
)
SELECT ID,YourString
,(
SELECT Chr As [*]
FROM SeparatedCharacters sc1
WHERE sc1.ID=t.ID
ORDER BY sc1.Chr
FOR XML PATH(''),TYPE
).value('.','nvarchar(max)') AS Sorted
FROM #tbl t;
The result
ID YourString Sorted
1 ABC ABC
2 JSKEzXO EJKOSXz
3 QKEvYUJMKRC CEJKKMQRUvY
The idea in short
The trick is the first CROSS APPLY. This will create a tally on-the-fly. You will get a resultset with numbers from 1 to n where n is the length of the current string.
The second apply uses this number to get each character one-by-one using SUBSTRING().
The outer SELECT calls from the orginal table, which means one-row-per-ID and use a correalted sub-query to fetch all related characters. They will be sorted and re-concatenated using FOR XML. You might add DISTINCT in order to avoid repeating characters.
That's it :-)
Hint: SQL-Server 2017+
With version v2017 there's the new function STRING_AGG(). This would make the re-concatenation very easy:
WITH SeparatedCharacters AS
(
SELECT *
FROM #tbl
CROSS APPLY
(SELECT TOP(LEN(YourString)) ROW_NUMBER() OVER(ORDER BY (SELECT NULL)) FROM master..spt_values) A(Nmbr)
CROSS APPLY
(SELECT SUBSTRING(YourString,Nmbr,1))B(Chr)
)
SELECT ID,YourString
,STRING_AGG(sc.Chr,'') WITHIN GROUP(ORDER BY sc.Chr) AS Sorted
FROM SeparatedCharacters sc
GROUP BY ID,YourString;
Considering your table having good amount of rows (~4 Million), I would suggest you to create a persisted calculated field in the table, to store these values. As calculating these values at run time in a view, will lead to performance problems.
If you are not able to normalize, add this as a denormalized column to the existing table.
I think the error you are getting could be due to empty codes.
If LEN(#str) = 0
BEGIN
SET #output = ''
END
ELSE
BEGIN
... EXISTING CODE BLOCK ...
END
I can suggest to split string into its characters using referred SQL function.
Then you can concatenate string back, this time ordered alphabetically.
Are you using SQL Server 2017? Because with SQL Server 2017, you can use SQL String_Agg string aggregation function to concatenate characters splitted in an ordered way as follows
select
t.CustId, string_agg(strval, '') within GROUP (order by strval)
from CharacterTable t
cross apply dbo.SPLIT(t.code) s
where strval is not null
group by CustId
order by CustId
If you are not working on SQL2017, then you can follow below structure using SQL XML PATH for concatenation in SQL
select
CustId,
STUFF(
(
SELECT
'' + strval
from CharacterTable ct
cross apply dbo.SPLIT(t.code) s
where strval is not null
and t.CustId = ct.CustId
order by strval
FOR XML PATH('')
), 1, 0, ''
) As concatenated_string
from CharacterTable t
order by CustId
The following query returns the results shown below:
SELECT
ProjectID, newID.value
FROM
[dbo].[Data] WITH(NOLOCK)
CROSS APPLY
STRING_SPLIT([bID],';') AS newID
WHERE
newID.value IN ('O95833', 'Q96NY7-2')
Results:
ProjectID value
---------------------
2 Q96NY7-2
2 O95833
2 O95833
2 Q96NY7-2
2 O95833
2 Q96NY7-2
4 Q96NY7-2
4 Q96NY7-2
Using the newly added STRING_AGG function (in SQL Server 2017) as it is shown in the following query I am able to get the result-set below.
SELECT
ProjectID,
STRING_AGG( newID.value, ',') WITHIN GROUP (ORDER BY newID.value) AS
NewField
FROM
[dbo].[Data] WITH(NOLOCK)
CROSS APPLY
STRING_SPLIT([bID],';') AS newID
WHERE
newID.value IN ('O95833', 'Q96NY7-2')
GROUP BY
ProjectID
ORDER BY
ProjectID
Results:
ProjectID NewField
-------------------------------------------------------------
2 O95833,O95833,O95833,Q96NY7-2,Q96NY7-2,Q96NY7-2
4 Q96NY7-2,Q96NY7-2
I would like my final output to have only unique elements as below:
ProjectID NewField
-------------------------------
2 O95833, Q96NY7-2
4 Q96NY7-2
Any suggestions about how to get this result? Please feel free to refine/redesign from scratch my query if needed.
Use the DISTINCT keyword in a subquery to remove duplicates before combining the results: SQL Fiddle
SELECT
ProjectID
,STRING_AGG(value, ',') WITHIN GROUP (ORDER BY value) AS
NewField
from (
select distinct ProjectId, newId.value
FROM [dbo].[Data] WITH(NOLOCK)
CROSS APPLY STRING_SPLIT([bID],';') AS newID
WHERE newID.value IN ( 'O95833' , 'Q96NY7-2' )
) x
GROUP BY ProjectID
ORDER BY ProjectID
You can use distinct in the subquery used for the apply:
SELECT d.ProjectID,
STRING_AGG( newID.value, ',') WITHIN GROUP (ORDER BY newID.value) AS
NewField
FROM [dbo].[Data] d CROSS APPLY
(select distinct value
from STRING_SPLIT(d.[bID], ';') AS newID
) newID
WHERE newID.value IN ( 'O95833' , 'Q96NY7-2' )
group by projectid;
This is a function that I wrote that answers the OP Title:
Improvements welcome!
CREATE OR ALTER FUNCTION [dbo].[fn_DistinctWords]
(
#String NVARCHAR(MAX)
)
RETURNS NVARCHAR(MAX)
WITH SCHEMABINDING
AS
BEGIN
DECLARE #Result NVARCHAR(MAX);
WITH MY_CTE AS ( SELECT Distinct(value) FROM STRING_SPLIT(#String, ' ') )
SELECT #Result = STRING_AGG(value, ' ') FROM MY_CTE
RETURN #Result
END
GO
Use like:
SELECT dbo.fn_DistinctWords('One Two Three Two One');
As #SeanLange pointed out in the comments, this is a terrible way to pull out the data, but if you had to, just make it 2 separate queries as follows:
SELECT
ProjectID
,STRING_AGG( val, ',') WITHIN GROUP (ORDER BY val) AS NewField
FROM
(
SELECT DISTINCT
ProjectID
,newID.value AS val
FROM
[dbo].[Data] WITH(NOLOCK)
CROSS APPLY STRING_SPLIT([bID],';') AS newID
WHERE
newID.value IN ('O95833' , 'Q96NY7-2')
) t
GROUP BY
ProjectID
That should do it.
Another possibility to get unique strings from STRING_AGG would be to perform these three steps after fetching the comma separated string:
Split the string (STRING_SPLIT)
Select DISTINCT from the splits
Apply STRING_AGG again to a select with a group on a single key
Example:
(select STRING_AGG(CAST(value as VARCHAR(MAX)), ',')
from (SELECT distinct 1 single_key, value
FROM STRING_SPLIT(STRING_AGG(CAST(customer_division as VARCHAR(MAX)), ','), ','))
q group by single_key) as customer_division
Here is my improvement on #ttugates to make it more generic:
CREATE OR ALTER FUNCTION [dbo].[fn_DistinctList]
(
#String NVARCHAR(MAX),
#Delimiter char(1)
)
RETURNS NVARCHAR(MAX)
WITH SCHEMABINDING
AS
BEGIN
DECLARE #Result NVARCHAR(MAX);
WITH MY_CTE AS ( SELECT Distinct(value) FROM STRING_SPLIT(#String,
#Delimiter) )
SELECT #Result = STRING_AGG(value, #Delimiter) FROM MY_CTE
RETURN #Result
END
You can make a distinct view of the table, that holds the aggregate values, that is even simpler:
Create Table Test (field1 varchar(1), field2 varchar(1));
go
Create View DistinctTest as (Select distinct field1, field2 from test group by field1,field2);
go
insert into Test Select 'A', '1';
insert into Test Select 'A', '2';
insert into Test Select 'A', '2';
insert into Test Select 'A', '2';
insert into Test Select 'D', '1';
insert into Test Select 'D', '1';
select string_agg(field1, ',') from Test where field2 = '1'; /* duplicates: A,D,D */;
select string_agg(field1, ',') from DistinctTest where field2 = '1'; /* no duplicates: A,D */;
Oracle (since version 19c) suports listagg (DISTINCT ..., but Microsoft SQL Server not probably.
I have a string coming like '1234_XXXX_RHL_PQR' and in output I want to pull character coming after 'XXXX_' that is 'RHL'.
I would always have 'XXXX' in string and my task is to get string after 'XXXX_'
Try This
DECLARE #Table AS TABLE (Data Nvarchar(100))
INSERT INTO #Table
SELECT '1234_XXXX_RHL_PQR'
SELECT Data,CAST('<S>'+REPLACE(Data,'_','</S><S>')+'</S>' AS XML).value('/S[3]','nvarchar(1000)') AS ReqData
FROM #Table
Result
Data ReqData
----------------------------
1234_XXXX_RHL_PQR RHL
DEMO : http://rextester.com/PXHSDZ80426
Try the following:
select
right(yourfield,len(yourfield)-3-patindex('%XXXX%',yourfield))
from yourtable
I would use substring() with charindex() function
select col, substring(col, charindex('XXXX_', col)+5, len(col)) as Ncol
from table t;
First Create Table-Value-Function
CREATE FUNCTION [dbo].[Udf_GetReqString](#IputData Varchar(200))
RETURNS #OutTable TABLE
(
ActualData varchar(200),
ReqData varchar(200)
)
AS
BEGIN
DECLARE #Table AS TABLE (Data Nvarchar(100))
INSERT INTO #Table
SELECT #IputData
INSERT INTO #OutTable
SELECT Data,SUBSTRING(DatReq,-1,CHARINDEX('_',Data )) AS DatReq
FROM
(
SELECT Data,SUBSTRING(Data,CHARINDEX('X_',Data )+2,LEN(Data) ) AS DatReq
FROM #Table
)dt
RETURN
END
Sample And Call Function
DECLARE #Table AS TABLE (Data Nvarchar(100))
INSERT INTO #Table
SELECT '1234_XXXX_SHL_PQR' UNION ALL
SELECT '1234_XXXX_RHL_PQR' UNION ALL
SELECT '1234_DDHDFD_XXXX_PHL_PQR' UNION ALL
SELECT '1234_ABAD_DADADA_XXXX_GHL_PQR'
SELECT t.Data,
udf.ReqData
FROM #Table t
CROSS APPLY [dbo].[Udf_GetReqString] (t.Data) As udf
Result
Data ReqData
-------------------------------------------
1234_XXXX_SHL_PQR SHL
1234_XXXX_RHL_PQR RHL
1234_DDHDFD_XXXX_PHL_PQR PHL
1234_ABAD_DADADA_XXXX_GHL_PQR GHL
DEMO
http://sqlfiddle.com/#!18/2bcfa/1