Hi!
Today I learned for xml path technique to concatenate strings in mssql. Since I've never worked with xml in mssql and google hasn't helped, I need to ask you.
Let's imagine the default case. We need to concatenate some strings from a table:
declare #xmlRepNames xml = (
select
', [' + report_name + ']'
from (
select distinct
report_order,
report_name
from #report
) x
order by
report_order
for xml path(''), type)
select
stuff((select #xmlRepNames.value('.', 'nvarchar(max)')), 1, 1, '')
So I get smth like this:
[str1], [str2], [strn]
Ok. It works fine. But I have two very similar concatenate blocks. The difference is just in the way the result string looks like:
[str1], [str2], [strn]
and
isnull([str1], 0) as [str1], isnull([str2], 0) as [str2], isnull([strn], 0) as [strn]
So I can write 2 very similar code blocks (already done, btw) with different string constructors or to try extend previous code to has xml variable containing 2 kind of constructors and then concatenate by xml node type. Doing 2nd way I met some problems - I wrote this:
declare #xmlRepNames xml = (
select
', [' + report_name + ']' as name,
', isnull([' + report_name + '], 0) as [' + report_name + ']' as res
from (
select distinct
report_order,
report_name
from #report
) x
order by
report_order
for xml path(''), type)
select
stuff((select #xmlRepNames.value('/name', 'nvarchar(max)')), 1, 1, ''),
stuff((select #xmlRepNames.value('/res', 'nvarchar(max)')), 1, 1, '')
but it raise error "XQuery [value()]: 'value()' requires a singleton (or empty sequence), found operand of type 'xdt:untypedAtomic *'".
To replace, e.g., '/name' to '/name[1]' or any other '/name[x]', will return just x-th 'name' record but not all 'name' records concatenated.
[question]: is it possible to solve problem 2nd way like I want and if it's possible then how?
[disclaimer]: the problem isn't really serious for me now (1st way just a little bit uglier but also fine), but it seems very interesting how to come over :)
Thanks!
Your subquery cannot return two values. If you just want to concatenate strings, you do not need the xml data type at all. You can do the stuff() and subquery in a single statement:
declare #Rep1Names nvarchar(max) = (
stuff((select ', [' + report_name + ']' as name
from (select distinct report_order, report_name
from #report
) x
order by report_order
for xml path('')
)
), 1, 1, '');
declare #Rep2Names nvarchar(max) = (
stuff(select ', isnull([' + report_name + '], 0) as [' + report_name + ']' as res
from (select distinct report_order, report_name
from #report
) x
order by report_order
for xml path('')
)
), 1, 1, '');
Ok, so I haven't been completely satisfied the way Gordon Linoff suggested and since I've found this kind of problem actual for me I'm adding here another solution without using for xml path:
declare
#pivot_sequence nvarchar(max),
#columns_sequence nvarchar(max)
select
#pivot_sequence = coalesce(#pivot_sequence + ', [', '[')
+ col_name + ']',
#columns_sequence = coalesce(#columns_sequence + ', ', '')
+ 'isnull([' + col_name + '], 0) as [' + col_name + ']'
from some_table
order by
/* some_columns if needed to order concatenation */
Obviously, it'll work much slower but in cases when you haven't many rows it won't drastically affect on performance. In my case I have dynamic pivot query and these strings are built for it - I won't have many columns in pivot so it's fine for me.
Related
We are using this small recursive script to create a string for the dynamic headers of a table - the variable feeds into a piece of dynamic sql.
DROP TABLE IF EXISTS #PivotStep1;
CREATE TABLE #PivotStep1(PlayMonth VARCHAR(25))
INSERT INTO #PivotStep1
values
('01-oct-2016'),
('01-nov-2016'),
('01-dec-2016'),
('01-jan-2017'),
('01-feb-2017');
DECLARE #ColumnName AS nvarchar(MAX)
SELECT #ColumnName = ISNULL(#ColumnName + ',','')
+ QUOTENAME(PlayMonth)
FROM (
SELECT DISTINCT PlayMonth
FROM #PivotStep1
) AS ps
SELECT #ColumnName;
The above gives the following:
[01-dec-2016],[01-feb-2017],[01-jan-2017],[01-nov-2016],[01-oct-2016]
What we want is to surround each column name with ISNULL and AS... So the desired output is the following:
ISNULL([01-dec-2016],0) AS [01-dec-2016],ISNULL([01-feb-2017],0) AS [01-feb-2017],ISNULL([01-jan-2017],0) AS [01-jan-2017],ISNULL([01-nov-2016],0) AS [01-nov-2016],ISNULL([01-oct-2016],0) AS [01-oct-2016]
We would like to keep using the above recursive approach rather than using STUFF and XML.
Use the quotename twice and do other required concatenation.
Try this:
select #ColumnName = ISNULL(#ColumnName + ',', '')
+ 'ISNULL(' + QUOTENAME(PlayMonth) + ', 0) as ' + QUOTENAME(PlayMonth)
from (
select distinct PlayMonth
from #PivotStep1
) as ps
select #ColumnName;
Demo
I am trying to figure out the result of the below query on how to concat all the columns into a single string with comma separated values. The number of values can be dynamic.
SELECT 'Sc4','Sc5','Sc8','Sc7','Sc2'
I want the result to be as below:
Sc4,Sc5,Sc8,Sc7,Sc2
I have tried using stuff but did manage to concat the string but unable to insert comma in between.
Below is what i have tried
SELECT Stuff((SELECT 'Sc4','Sc5','Sc8','Sc7','Sc2' FOR XML PATH('')), 2, 0, '');
UPDATE
I would like to rephrase my question, is there any way the out shown below can be converted to csv
i assume Sc4, Sc5 etc are your column name and not string constant ?
SELECT Stuff(
(
SELECT ',' + Sc4 + ',' + Sc5 + ',' + Sc8 + ',' + Sc7
+ ',' + Sc2
FROM yourtable
FOR XML PATH('')
), 1, 1, '');
Try concatenation like
SELECT 'Sc4' + ',' + 'Sc5' + ',' + 'Sc8' + ',' + 'Sc7' + ',' + 'Sc2'
I have a table variable in a stored procedure. What I want is to find all of the unique values in one column and join them in a comma-separated list. I am already in a stored procedure, so I can do it some way that way; however, I am curious if I can do this with a query. I am on SQL Server 2008. This query gets me the values I want:
SELECT DISTINCT faultType FROM #simFaults;
Is there a way (using CONCAT or something like that) where I can get the list as a single comma-separated value?
This worked for me on a test dataset.
DECLARE #MyCSV Varchar(200) = ''
SELECT #MyCSV = #MyCSV +
CAST(faulttype AS Varchar) + ','
FROM #Simfaults
GROUP BY faultType
SET #MyCSV = LEFT(#MyCSV, LEN(#MyCSV) - 1)
SELECT #MyCSV
The last part is needed to trim the trailing comma.
+1 to JNK - the other common way you will see, which doesn't require a variable is:
SELECT DISTINCT faulttype + ','
FROM #simfaults
FOR XML PATH ('')
Note that if faulttype contains characters like "<" for example, those will be xml encoded. But for simple values this will be OK.
this is how we do this
create table #test (item int)
insert into #test
values(1),(2),(3)
select STUFF((SELECT ', ' + cast(Item as nvarchar)
FROM #test
FOR XML PATH('')), 1, 2, '')
Without the space after the comma it would be;
select STUFF((SELECT ',' + cast(Item as nvarchar)
FROM #test
FOR XML PATH('')), 1,1, '')
I am using FOR XML in a query to join multiple rows together, but the text contains quotes, "<", ">", etc. I need the actual character instead of the encoded value like """ etc. Any suggestions?
Basically what you're asking for is invalid XML and luckly SQL Server will not produce it. You can take the generated XML and extract the content, and this operation will revert the escaped characters to their text representation. This revert normally occurs in the presnetaitonlayer, but it can occur in SQL Server itslef by instance using XML methods to extract the content of the produced FOR XML output. For example:
declare #text varchar(max) = 'this text has < and >';
declare #xml xml;
set #xml = (select #text as [node] for xml path('nodes'), type);
select #xml;
select x.value(N'.', N'varchar(max)') as [text]
from #xml.nodes('//nodes/node') t(x);
I have a similar requirement to extract column names for use in PIVOT query.
The solution I used was as follows:
SELECT #columns = STUFF((SELECT '],[' + Value
FROM Table
ORDER BY Value
FOR XML PATH('')), 1, 2, '') + ']'
This produces a single string:
[Value 1],[Value 2],[Value 3]
I hope this points you in the right direction.
--something like this?
SELECT * INTO #Names FROM (
SELECT Name='<>&' UNION ALL
SELECT Name='ab<>'
) Names;
-- 1)
SELECT STUFF(
(SELECT ', ' + Name FROM #Names FOR XML PATH(''))
,1,2,'');
-- 2)
SELECT STUFF(
(SELECT ', ' + Name FROM #Names FOR XML PATH(''),TYPE).value('text()[1]','nvarchar(max)')
,1,2,'');
-- 2) is slower but will not return encoded value.
Hope it help.
I'm working on a dynamic pivot query on a table that contains:
OID - OrderID
Size - size of the product
BucketNum - the order that the sizes
should go
quantity - how many ordered
The size column contains different sizes depending upon the OID.
So, using the code found here, I put this together:
DECLARE #listCol VARCHAR(2000)
DECLARE #query VARCHAR(4000)
SELECT #listCol = STUFF(( SELECT distinct '], [' + [size]
FROM #t
FOR
XML PATH('')
), 1, 2, '') + ']'
SET #query = 'SELECT * FROM
(SELECT OID, [size], [quantity]
FROM #t
) src
PIVOT (SUM(quantity) FOR Size
IN (' + #listCol + ')) AS pvt'
EXECUTE ( #query )
This works great except that the column headers (the sizes labels) are not in the order based upon the bucketnum column. The are in the order based upon the sizes.
I've tried the optional Order By after the pivot, but that is not working.
How do I control the order in which the columns appear?
Thank you
You need to fix this:
SELECT #listCol = STUFF(( SELECT distinct '], [' + [size]
FROM #t
FOR
XML PATH('')
), 1, 2, '') + ']'
To return the columns in the right order. You might have to do something like this instead of using DISTINCT:
SELECT [size]
FROM #t
GROUP BY [size]
ORDER BY MIN(BucketNum)
SELECT #listCol = STUFF(
(SELECT DISTINCT ',' + QUOTENAME(size) AS [size]
FROM #t
ORDER BY [size]
FOR XML PATH('')
I saw this link just today, which uses a CTE to build the column list (which, presumably, you could order) on the fly without the need for dynamic sql:
http://blog.stevienova.com/2009/07/13/using-ctes-to-create-dynamic-pivot-tables-in-sql-20052008/
I had the same problem and tried the solution suggested above but, probably due to my level of understanding, couldn't get it to work. I found a simple hack was to create a Temp table with the column headers ordered correctly using Order by statements and then pull in that list to the variable that sets the dynamic pivot query column names.
e.g.
SELECT WeekNum INTO #T3
FROM #T2
GROUP BY WeekNum
ORDER BY MIN(WeekNum)
SELECT #ColumnName1 = ISNULL(#ColumnName1 + ',','') + QuoteName(WeekNum)
FROM (SELECT WeekNum From #T3) AS WeekNum
Worked a treat.
Hope that helps someone.