SQL query with subquery and concatenating variable using CRM on-premise data - sql

I am working on a report where I need to provide a summary of notes for particular "activities/tasks".
Since the activity can accept multiple notes, I have to search for all the notes related to that activity. I then order it by date (new to old), and concatenate them with some other strings as such:
[Tom Smith wrote on 9/23/2016 1:21 pm] Client was out of office, left message. [Jane Doe wrote on 9/21/2016 3:24 pm] Client called asking about pricing.
The data comes from replicated tables of our on-premise CRM system, and I'm using SQL Server 2012. The tables I'm using are: AnnotationBase (contains the notes), ActivityPointerBase (contains the activities/tasks), and SystemUserID (to lookup usernames). Due to Data mismatch, I have to do some converting of the data types so that I can concatenate them properly, so that's why there's a lot of CAST and CONVERT. In addition, not all Activities have a NoteText associated with them, and sometimes the NoteText field is NULL, so I have to catch and filter the NULLs out (or it'll break my concatenated string).
I have written the following query:
DECLARE #Notes VarChar(Max)
SELECT
( SELECT TOP 5 #Notes = COALESCE(#Notes+ ', ', '') + '[' + CONVERT(varchar(max), ISNULL(sUB.FullName, 'N/A')) + ' wrote on ' + CONVERT(varchar(10), CAST(Anno.ModifiedOn AS DATE), 101) + RIGHT(CONVERT(varchar(32),Anno.ModifiedOn,100),8) + '] ' + CONVERT(varchar(max), ISNULL(Anno.NoteText, '')) --+ CONVERT(varchar(max), CAST(ModifiedOn AS varchar(max)), 101)--+ CAST(ModifiedOn AS varchar(max))
FROM [CRM_rsd].[dbo].[AnnotationBase] AS Anno
LEFT OUTER JOIN [CRM_rsd].[dbo].[systemUserBase] AS sUB
ON Anno.ModifiedBy = sUB.SystemUserId
WHERE Anno.ObjectId = Task.ActivityId--'0B48AB28-C08F-419A-8D98-9916BDFFDE4C'
ORDER BY Anno.ModifiedOn DESC
SELECT LEFT(#Notes,LEN(#Notes)-1)
) AS Notes
,Task.*
FROM [CRM_rsd].[dbo].[ActivityPointerBase] AS Task
WHERE Task.Subject LIKE '%Project On Hold%'
I know the above method is probably not very efficient, but the list of "Projects On Hold" is rather small (less than 500), so performance isn't a priority. What is a priority is to be able to get a consolidated and concatenated list of notes for each activity. I have been searching all over the internet for a solution, and I have tried many different methods. But I get the following errors:
Msg 102, Level 15, State 1, Line 3
Incorrect syntax near '='.
Msg 102, Level 15, State 1, Line 10
Incorrect syntax near ')'.
I envision two possible solutions to my problem:
My subquery errors are fixed or
Create a "view" of just the concatenated NotesText, grouped by ActivityId (which would work as a key), and then just query from that.
Yet even though I'm pretty sure my ideas would work, I can't seem to figure out how to concatenate a column and group at the same time.

What you are trying to do is display the records from one table (in your case ActivityPointerBase) and inside you want to add a calculated column with information from multiple records from another table (in your case AnnotationBase) merged in the rows.
There are multiple ways how you could achieve this that are different in terms of performance impact:
Option 1. You could write a scalar function that would receive as parameter the id of the task and would inside select the top 5 records, concatenating them in a procedural fashion and returning a varchar(max)
Option 2: You could use a subquery in combination with the FOR XML clause.
SELECT
SUBSTRING(
CAST(
(SELECT TOP 5 ', ' +
'[' + CONVERT(varchar(max), ISNULL(FullName, 'N/A')) +
' wrote on ' +
CONVERT(varchar(10), CAST(ModifiedOn AS DATE), 101) +
RIGHT(CONVERT(varchar(32),ModifiedOn,100),8) + '] ' +
CONVERT(varchar(max), ISNULL(NoteText, ''))
FROM [CRM_rsd].[dbo].[AnnotationBase] AS Anno
LEFT OUTER JOIN [CRM_rsd].[dbo].[systemUserBase] AS sUB ON Anno.ModifiedBy = sUB.SystemUserId
WHERE Anno.ObjectId = Task.ActivityId
ORDER BY Anno.ModifiedOn DESC
FOR XML PATH(''),TYPE
) AS VARCHAR(MAX)
),3,99999) AS Notes
,Task.*
FROM [CRM_rsd].[dbo].[ActivityPointerBase] AS Task
WHERE Task.Subject LIKE '%Project On Hold%'
What here happens is that by using the construct inside the CAST() we fetch the top 5 lines and make SQL server produce an XML with no element names, resulting in concatenation of the element values, we add comma as separator. Then we convert the XML to varchar(max) and remove the initial separator before the first record.
I prefer option 2, it will perform much better then using a scalar function.

Related

Is there a way to split text from cell into multiple records in SQL using line break delimiter?

I have a system generated report that exports all expense reports for a given period. All employees attached to the report are exported in one field "Attendees", separating multiple employees with line breaks.
I would like to break the cell into multiple records, still including the expense report NO. on each record.
Note, the names in the attendee field are different formats, but they all end with a ")".
Is there a way to use the line break as a delimiter and create a new record with each break, and populate with the same values of the other fields in the original record?
SQL would be the ideal interface, but I can use Python, R, Tableau, and PowerBI before importing into the server if need be.
Example below.
&
SQL Server:
SELECT ExpRep, ltrim(value+')')
FROM employees
CROSS APPLY STRING_SPLIT(Attendees, ')')
WHERE RTRIM(value) <> '';
Oracle:
select ExpRep
, Attendees || ')'
from (select ExpRep
, ltrim(regexp_substr(Attendees, '[^)]+', 1, levels.column_value)) as Attendees
from test
, table(cast(multiset(select level
from dual connect by level <= length (regexp_replace(Attendees, '[^)]+')) + 1) as sys.OdciNumberList)) levels)
where Attendees is not null;
MySQL:
(Only valid for examples where you will have 5 different employees in column Attendees. If yes then add union all...)
select
ExpRep,
concat(ltrim(SUBSTRING_INDEX(SUBSTRING_INDEX(Attendees, ')', numbers.n), ')', -1)),')') name
from (select 1 n
union all select 2
union all select 3
union all select 4
union all select 5) numbers INNER JOIN test
on CHAR_LENGTH(Attendees)
-CHAR_LENGTH(REPLACE(Attendees, ')', ''))>=numbers.n-1
where ltrim(SUBSTRING_INDEX(SUBSTRING_INDEX(Attendees, ')', numbers.n), ')', -1)) <> ''
order by
ExpRep, n
Note
It would be nice that one of this will be of use to you but next time do add a tag of a database you use.

Using Upper to Capitalize the first letter of City name

I am doing some data clean-up and need to Capitalize the first letter of City names. How do I capitalize the second word in a City Like Terra Bella.
SELECT UPPER(LEFT([MAIL CITY],1))+
LOWER(SUBSTRING([MAIL CITY],2,LEN([MAILCITY])))
FROM masterfeelisting
My results is this 'Terra bella' and I need 'Terra Bella'. Thanks in advance.
Ok, I know I answered this before, but it bugged me that we couldn't write something efficient to handle an unknown amount of 'text segments'.
So re-thinking it and researching, I discovered a way to change the [MAILCITY] field into XML nodes where each 'text segment' is assigned it's own Node within the xml field. Then those xml fields can be processed node by node, concatenated together, and then changed back to a SQL varchar. It's convoluted, but it works. :)
Here's the code:
CREATE TABLE
#masterfeelisting (
[MAILCITY] varchar(max) not null
);
INSERT INTO #masterfeelisting VALUES
('terra bellA')
,(' terrA novA ')
,('chicagO ')
,('bostoN')
,('porT dE sanTo')
,(' porT dE sanTo pallo ');
SELECT
RTRIM
(
(SELECT
UPPER([xmlField].[xmlNode].value('.', 'char(1)')) +
LOWER(STUFF([xmlField].[xmlNode].value('.', 'varchar(max)'), 1, 1, '')) + ' '
FROM [xmlNodeRecordSet].[nodeField].nodes('/N') as [xmlField]([xmlNode]) FOR
xml path(''), type
).value('.', 'varchar(max)')
) as [MAILCITY]
FROM
(SELECT
CAST('<N>' + REPLACE([MAILCITY],' ','</N><N>')+'</N>' as xml) as [nodeField]
FROM #masterfeelisting
) as [xmlNodeRecordSet];
Drop table #masterfeelisting;
First I create a table and fill it with dummy values.
Now here is the beauty of the code:
For each record in #masterfeelisting, we are going to create an xml field with a node for each 'text segment'.
ie. '<N></N><N>terrA</N><N>novA</N><N></N>'
(This is built from the varchar ' terrA novA ')
1) The way this is done is by using the REPLACE function.
The string starts with a '<N>' to designate the beginning of the node. Then:
REPLACE([MAILCITY],' ','</N><N>')
This effectively goes through the whole [MAILCITY] string and replaces each
' ' with '</N><N>'
and then the string ends with a '</N>'. Where '</N>' designates the end of each node.
So now we have a beautiful XML string with a couple of empty nodes and the 'text segments' nicely nestled in their own node. All the 'spaces' have been removed.
2) Then we have to CAST the string into xml. And we will name that field [nodeField]. Now we can use xml functions on our newly created record set. (Conveniently named [xmlNodeRecordSet].)
3) Now we can read the [xmlNodeRecordSet] into the main sub-Select by stating:
FROM [xmlNodeRecordSet].[nodeField].nodes('/N')
This tells us we are reading the [nodeField] as nodes with a '/N' delimiter.
This table of node fields is then parsed by stating:
as [xmlField]([xmlNode]) FOR xml path(''), type
This means each [xmlField] will be parsed for each [xmlNode] in the xml string.
4) So in the main sub-select:
Each blank node '<N></N>' is discarded. (Or not processed.)
Each node with a 'text segment' in it will be parsed. ie <N>terrA</N>
UPPER([xmlField].[xmlNode].value('.', 'char(1)')) +
This code will grab each node out of the field and take its contents '.' and only grab the first character 'char(1)'. Then it will Upper case that character. (the plus sign at the end means it will concatenate this letter with the next bit of code:
LOWER(STUFF([xmlField].[xmlNode].value('.', 'varchar(max)'), 1, 1, ''))
Now here is the beauty... STUFF is a function that will take a string, from a position, for a length, and substitute another string.
STUFF(string, start position, length, replacement string)
So our string is:
[xmlField].[xmlNode].value('.', 'varchar(max)')
Which grabs the whole string inside the current node since it is 'varchar(max)'.
The start position is 1. The length is 1. And the replacement string is ''. This effectively strips off the first character by replacing it with nothing. So the remaining string is all the other characters that we want to have lower case. So that's what we do... we use LOWER to make them all lower case. And this result is concatenated to our first letter that we already upper cased.
But wait... we are not done yet... we still have to append a + ' '. Which adds a blank space after our nicely capitalized 'text segment'. Just in case there is another 'text segment' after this node is done.
This main sub-Select will now parse each node in our [xmlField] and concatenate them all nicely together.
5) But now that we have one big happy concatenation, we still have to change it back from an xml field to a SQL varchar field. So after the main sub-select we need:
.value('.', 'varchar(max)')
This changes our [MAILCITY] back to a SQL varchar.
6) But hold on... we still are not done. Remember we put an extra space at the end of each 'text segment'??? Well the last 'text segment still has that extra space after it. So we need to Right Trim that space off by using RTRIM.
7) And dont forget to rename the final field back to as [MAILCITY]
8) And that's it. This code will take an unknown amount of 'text segments' and format each one of them. All using the fun of XML and it's node parsers.
Hope that helps :)
Here's one way to handle this using APPLY. Note that this solution supports up to 3 substrings (e.g. "Phoenix", "New York", "New York City") but can easily be updated to handle more.
DECLARE #string varchar(100) = 'nEW yoRk ciTY';
WITH DELIMCOUNT(String, DC) AS
(
SELECT #string, LEN(RTRIM(LTRIM(#string)))-LEN(REPLACE(RTRIM(LTRIM(#string)),' ',''))
),
CIPOS AS
(
SELECT *
FROM DELIMCOUNT
CROSS APPLY (SELECT CHARINDEX(char(32), string, 1)) CI1(CI1)
CROSS APPLY (SELECT CHARINDEX(char(32), string, CI1.CI1+1)) CI2(CI2)
)
SELECT
OldString = #string,
NewString =
CASE DC
WHEN 0 THEN UPPER(SUBSTRING(string,1,1))+LOWER(SUBSTRING(string,2,8000))
WHEN 1 THEN UPPER(SUBSTRING(string,1,1))+LOWER(SUBSTRING(string,2,CI1-1)) +
UPPER(SUBSTRING(string,CI1+1,1))+LOWER(SUBSTRING(string,CI1+2,100))
WHEN 2 THEN UPPER(SUBSTRING(string,1,1))+LOWER(SUBSTRING(string,2,CI1-1)) +
UPPER(SUBSTRING(string,CI1+1,1))+LOWER(SUBSTRING(string,CI1+2,CI2-(CI1+1))) +
UPPER(SUBSTRING(string,CI2+1,1))+LOWER(SUBSTRING(string,CI2+2,100))
END
FROM CIPOS;
Results:
OldString NewString
--------------- --------------
nEW yoRk ciTY New York City
This will only capitalize the first letter of the second word. A shorter but less flexible approach. Replace #str with [Mail City].
DECLARE #str AS VARCHAR(50) = 'Los angelas'
SELECT STUFF(#str, CHARINDEX(' ', #str) + 1, 1, UPPER(SUBSTRING(#str, CHARINDEX(' ', #str) + 1, 1)));
This is a way to use imbedded Selects for three City name parts.
It uses CHARINDEX to find the location of your separator character. (ie a space)
I put an 'if' structure around the Select to test if you have any records with more than 3 parts to the city name. If you ever get the warning message, you could add another sub-Select to handle another city part.
Although... just to be clear... SQL is not the best language to do complicated formatting. It was written as a data retrieval engine with the idea that another program will take that data and massage it into a friendlier look and feel. It may be easier to handle the formatting in another program. But if you insist on using SQL and you need to account for city names with 5 or more parts... you may want to consider using Cursors so you can loop through the variable possibilities. (But Cursors are not a good habit to get into. So don't do that unless you've exhausted all other options.)
Anyway, the following code creates and populates a table so you can test the code and see how it works. Enjoy!
CREATE TABLE
#masterfeelisting (
[MAILCITY] varchar(30) not null
);
Insert into #masterfeelisting select 'terra bella';
Insert into #masterfeelisting select ' terrA novA ';
Insert into #masterfeelisting select 'chicagO ';
Insert into #masterfeelisting select 'bostoN';
Insert into #masterfeelisting select 'porT dE sanTo';
--Insert into #masterfeelisting select ' porT dE sanTo pallo ';
Declare #intSpaceCount as integer;
SELECT #intSpaceCount = max (len(RTRIM(LTRIM([MAILCITY]))) - len(replace([MAILCITY],' ',''))) FROM #masterfeelisting;
if #intSpaceCount > 2
SELECT 'You need to account for more than 3 city name parts ' as Warning, #intSpaceCount as SpacesFound;
else
SELECT
cThird.[MAILCITY1] + cThird.[MAILCITY2] + cThird.[MAILCITY3] as [MAILCITY]
FROM
(SELECT
bSecond.[MAILCITY1] as [MAILCITY1]
,SUBSTRING(bSecond.[MAILCITY2],1,bSecond.[intCol2]) as [MAILCITY2]
,UPPER(SUBSTRING(bSecond.[MAILCITY2],bSecond.[intCol2] + 1, 1)) +
SUBSTRING(bSecond.[MAILCITY2],bSecond.[intCol2] + 2,LEN(bSecond.[MAILCITY2]) - bSecond.[intCol2]) as [MAILCITY3]
FROM
(SELECT
SUBSTRING(aFirst.[MAILCITY],1,aFirst.[intCol1]) as [MAILCITY1]
,UPPER(SUBSTRING(aFirst.[MAILCITY],aFirst.[intCol1] + 1, 1)) +
SUBSTRING(aFirst.[MAILCITY],aFirst.[intCol1] + 2,LEN(aFirst.[MAILCITY]) - aFirst.[intCol1]) as [MAILCITY2]
,CHARINDEX ( ' ', SUBSTRING(aFirst.[MAILCITY],aFirst.[intCol1] + 1, LEN(aFirst.[MAILCITY]) - aFirst.[intCol1]) ) as intCol2
FROM
(SELECT
UPPER (LEFT(RTRIM(LTRIM(mstr.[MAILCITY])),1)) +
LOWER(SUBSTRING(RTRIM(LTRIM(mstr.[MAILCITY])),2,LEN(RTRIM(LTRIM(mstr.[MAILCITY])))-1)) as [MAILCITY]
,CHARINDEX ( ' ', RTRIM(LTRIM(mstr.[MAILCITY]))) as intCol1
FROM
#masterfeelisting as mstr -- Initial Master Table
) as aFirst -- First Select Shell
) as bSecond -- Second Select Shell
) as cThird; -- Third Select Shell
Drop table #masterfeelisting;

How to select specific text from the string in efficient way using SQL

Below is my query. It is giving me correct output but I need to run it efficiently as it is used for 500k records.
DECLARE #DESC_MESSAGE VARCHAR(5000)
SET #DESC_MESSAGE = '12345 VENKAT was entered ODC ABCD-3'
SELECT REPLACE(#DESC_MESSAGE,SUBSTRING(#DESC_MESSAGE,1,CHARINDEX('was',#DESC_MESSAGE,3)-1),'')
I just want to retrieve text after 'was' which can change depending on condition.
for ex. text can be like
'112233 XYZ was entered ODC PQRS-3' or
'223344 HARRY was gone out of ODC AMD-3'
Please suggest efficient way to retrieve such text.
I would be inclined to use stuff():
select stuff(col, 1, chardindex('was ', col + 'was ') + 4, '')
The + 'was + in the charindex() function just guarantees no error if 'was ' is not in the text.
half milion rows is not so huge..
what i can see in your question is that there is an architecture issue,
why do you need to split a column to make a query?
why don't you keep the colums splitted in origin ?
eventually you could have another column that contains only the text after the "was"
this could be better even if the rows grow a lot.
select LTRIM(stuff(#DESC_MESSAGE, 1, CHARINDEX(' was', #DESC_MESSAGE + 'was') + 3, ''))

SQL Server bringing back inconsistent results

I am working on some reports that were created before I started in my current job. One of these reports is based on a view (SQL Server 2005).
This view is incredibly large and unwieldy, and for now, I won't post it because I think it's just too big. I'm not sure how it was produced - I'm guessing that it was produced in the designer because I can't see someone actually writing stuff like this. It's several pages long, and references 5 other views. Bottom line - it's complicated, and needs to be refactored/redesigned, but until we get time for that we're stuck with it.
Anyway, I have to make some minor non-functional changes to it in order to move it to a different database and schema. In order to make sure I'm not changing what it actually returns, I'm amending a second version of the view. Let's call the first view vw_Data1 and my new view vw_Data2. Now, if I write:
SELECT Count(*) FROM
(
SELECT * FROM vw_Data1
UNION
SELECT * FROM vw_Data2
)
then I should get back the same number as if I just did
SELECT Count(*) FROM vw_Data1
as long as vw_Data1 and vw_Data2 return identical rows (which is what I want to check).
However, what I am finding is if I run the UNION query above several times, I get DIFFERENT RESULTS EACH TIME.
So, just to be clear, if I run:
SELECT Count(*) FROM
(
SELECT * FROM vw_Data1
UNION
SELECT * FROM vw_Data2
)
more than once, then I get different results each time.
As I say, I'm not posting the actual code yet, because the first thing I want to ask is simply this - how on earth can a query return different results?
There is one non-deterministic function used, and that is as part of the following (horrible) join:
LEFT OUTER JOIN dbo.vwuniversalreportingdata_budget
ON
CASE
WHEN dbo.f_tasks.ta_category = 'Reactive' THEN
CAST(dbo.f_tasks.ta_fkey_fc_seq AS VARCHAR(10))
+ ' | '
+ CAST(dbo.f_tasks.ta_fkey_fcc_seq AS VARCHAR(10))
+ ' | '
+ CAST(YEAR(DATEADD(MONTH, -3, dbo.f_tasks.ta_sched_date)) AS VARCHAR(10))
WHEN dbo.f_tasks.ta_category = 'Planned' THEN
CAST(dbo.f_tasks.ta_fkey_fc_seq AS VARCHAR(10))
+ ' | '
+ CAST(dbo.f_tasks.ta_fkey_fcc_seq AS VARCHAR(10))
+ ' | '
+ CAST(YEAR(DATEADD(MONTH, -3, dbo.f_tasks.ta_est_date)) AS VARCHAR(10))
WHEN dbo.f_tasks.ta_category = 'Periodic' THEN
CAST(dbo.f_tasks.ta_fkey_fc_seq AS VARCHAR(10))
+ ' | '
+ CAST(dbo.f_tasks.ta_fkey_fcc_seq AS VARCHAR(10))
+ ' | '
+ CAST(YEAR(DATEADD(MONTH, -3, dbo.f_tasks.ta_est_date)) AS VARCHAR(10))
END
= dbo.vwuniversalreportingdata_budget.id
The whole query is pretty disgusting like this. Anyway, any thoughts on how this could happen would be gratefully received. Is it something to do with the union, perhaps? I don't know. Help!

SQL Query - Concatenating Results into One String [duplicate]

This question already has answers here:
How to concatenate text from multiple rows into a single text string in SQL Server
(47 answers)
Closed 7 years ago.
I have a sql function that includes this code:
DECLARE #CodeNameString varchar(100)
SELECT CodeName FROM AccountCodes ORDER BY Sort
I need to concatenate all results from the select query into CodeNameString.
Obviously a FOREACH loop in C# code would do this, but how do I do it in SQL?
If you're on SQL Server 2005 or up, you can use this FOR XML PATH & STUFF trick:
DECLARE #CodeNameString varchar(100)
SELECT
#CodeNameString = STUFF( (SELECT ',' + CodeName
FROM dbo.AccountCodes
ORDER BY Sort
FOR XML PATH('')),
1, 1, '')
The FOR XML PATH('') basically concatenates your strings together into one, long XML result (something like ,code1,code2,code3 etc.) and the STUFF puts a "nothing" character at the first character, e.g. wipes out the "superfluous" first comma, to give you the result you're probably looking for.
UPDATE: OK - I understand the comments - if your text in the database table already contains characters like <, > or &, then my current solution will in fact encode those into <, >, and &.
If you have a problem with that XML encoding - then yes, you must look at the solution proposed by #KM which works for those characters, too. One word of warning from me: this approach is a lot more resource and processing intensive - just so you know.
DECLARE #CodeNameString varchar(max)
SET #CodeNameString=''
SELECT #CodeNameString=#CodeNameString+CodeName FROM AccountCodes ORDER BY Sort
SELECT #CodeNameString
#AlexanderMP's answer is correct, but you can also consider handling nulls with coalesce:
declare #CodeNameString nvarchar(max)
set #CodeNameString = null
SELECT #CodeNameString = Coalesce(#CodeNameString + ', ', '') + cast(CodeName as varchar) from AccountCodes
select #CodeNameString
For SQL Server 2005 and above use Coalesce for nulls and I am using Cast or Convert if there are numeric values -
declare #CodeNameString nvarchar(max)
select #CodeNameString = COALESCE(#CodeNameString + ',', '') + Cast(CodeName as varchar) from AccountCodes ORDER BY Sort
select #CodeNameString
from msdn Do not use a variable in a SELECT statement to concatenate values (that is, to compute aggregate values). Unexpected query results may occur. This is because all expressions in the SELECT list (including assignments) are not guaranteed to be executed exactly once for each output row
The above seems to say that concatenation as done above is not valid as the assignment might be done more times than there are rows returned by the select
Here is another real life example that works fine at least with 2008 release (and later).
This is the original query which uses simple max() to get at least one of the values:
SELECT option_name, Field_M3_name, max(Option_value) AS "Option value", max(Sorting) AS "Sorted"
FROM Value_list group by Option_name, Field_M3_name
ORDER BY option_name, Field_M3_name
Improved version, where the main improvement is that we show all values comma separated:
SELECT from1.keys, from1.option_name, from1.Field_M3_name,
Stuff((SELECT DISTINCT ', ' + [Option_value] FROM Value_list from2
WHERE COALESCE(from2.Option_name,'') + '|' + COALESCE(from2.Field_M3_name,'') = from1.keys FOR XML PATH(''),TYPE)
.value('text()[1]','nvarchar(max)'),1,2,N'') AS "Option values",
Stuff((SELECT DISTINCT ', ' + CAST([Sorting] AS VARCHAR) FROM Value_list from2
WHERE COALESCE(from2.Option_name,'') + '|' + COALESCE(from2.Field_M3_name,'') = from1.keys FOR XML PATH(''),TYPE)
.value('text()[1]','nvarchar(max)'),1,2,N'') AS "Sorting"
FROM ((SELECT DISTINCT COALESCE(Option_name,'') + '|' + COALESCE(Field_M3_name,'') AS keys, Option_name, Field_M3_name FROM Value_list)
-- WHERE
) from1
ORDER BY keys
Note that we have solved all possible NULL case issues that I can think of and also we fixed an error that we got for numeric values (field Sorting).