January 13, 2009 at 7:42 am
Dear Members,
How to Convert nx1 dimension table to 1x1 dimension table
here n - rows and 1 - columns.
-------
For example,
DECLARE @tblVariable table(Roles varchar(2000))
insert into @tblVariable VALUES('Role1,Role2,')
insert into @tblVariable VALUES('Role1,Role3')
insert into @tblVariable VALUES('Role1,Role4')
insert into @tblVariable VALUES('Role2,Role5)
insert into @tblVariable VALUES('Role6,Role1)
insert into @tblVariable VALUES('Role7')
insert into @tblVariable VALUES('Role8)
SELECT * FROM @tblVariable
--Out Put of the select query
Role1,Role2,
Role1,Role3,
Role1,Role4,
Role2,Role5,
Role6,Role1,
Role7,
Role8,
--My final requierment is to get all distinct roles with comma separation.
--
-- Role1, Role2, Role3, Role4, Role5, Role6, Role7, Role8
actually the Roles in the table will be in Roles Tag (here roles table is just like html tags)
Please Help me in this scenario.
Thank you,
January 13, 2009 at 7:50 am
There are many articles on this forum about concatenating strings.
Here is an excellent article by Jeff Moden on it
http://www.sqlservercentral.com/articles/Test+Data/61572/%5B/url%5D%5B/u%5D
--Ramesh
January 13, 2009 at 8:37 am
yes you are rigth,
refernce to the link
-----------------
http://www.sqlservercentral.com/articles/Test+Data/61572/
Using function we can do,
CREATE FUNCTION dbo.fnConcatTest (@SomeID INT)RETURNS VARCHAR(8000) AS BEGIN DECLARE @Return VARCHAR(8000)
SELECT @Return = ISNULL(@Return+',','')+SomeCode FROM dbo.TestData WHERE SomeID = @SomeID ORDER BY RowNum
RETURN @Return
END
but my problem i want the query to get the result as "Role1, Role2, Role3..."
so i can subquery and filter all the records with these roles.
here i my scenario i am not using the stored procedure and function.
is it possible to achieve my requiremt without using function.
-------
Santosh
January 15, 2009 at 12:27 am
Hi All,
Thanks a lot for your help,
I have done this using Cursors,
Here is the complete code,
DECLARE @tblRolesWithComma TABLE(Roles VARCHAR(8000))
insert into @tblRolesWithComma VALUES('Role1,Role2,')
insert into @tblRolesWithComma VALUES('Role1,Role3')
insert into @tblRolesWithComma VALUES('Role1,Role4')
insert into @tblRolesWithComma VALUES('Role2,Role5')
insert into @tblRolesWithComma VALUES('Role6,Role1')
insert into @tblRolesWithComma VALUES('Role7')
insert into @tblRolesWithComma VALUES('Role8')
SELECT * FROM @tblRolesWithComma
DECLARE @tblRolesInSingleColumn TABLE(RoleId VARCHAR(8000))
DECLARE cur_Roles CURSOR
READ_ONLY
FOR (SELECT * FROM @tblRolesWithComma)
DECLARE @RoleIds VARCHAR(8000)
OPEN cur_Roles
declare @pos int
declare @RoleId VARCHAR(500)
FETCH NEXT FROM cur_Roles INTO @RoleIds
WHILE (@@fetch_status <> -1)
BEGIN
IF (@@fetch_status <> -2)
BEGIN
IF RIGHT(RTRIM(@RoleIds),1) <> ','
SET @RoleIds = @RoleIds + ','
SET @pos = PATINDEX('%,%' , @RoleIds)
WHILE @pos <> 0
BEGIN
SET @RoleId = left(@RoleIds, @pos - 1)
INSERT INTO @tblRolesInSingleColumn VALUES(CAST(@RoleId as varchar(500)))
SET @RoleIds = STUFF(@RoleIds, 1, @pos, '')
SET @pos = PATINDEX('%,%' , @RoleIds)
END
END
FETCH NEXT FROM cur_Roles INTO @RoleIds
END
CLOSE cur_Roles
DEALLOCATE cur_Roles
--SELECT DISTINCT * FROM @tblRolesInSingleColumn
DECLARE @Result varchar(8000)
SET @Result = ''
SELECT @Result = @Result + CASE WHEN LEN(@Result)>0 THEN ', ' ELSE '' END + RoleId
FROM (SELECT DISTINCT RoleId FROM @tblRolesInSingleColumn)Roles
SELECT @Result RolesIds
SET NOCOUNT OFF
GO
Thank you very much dudes for you help
-----
Santosh
January 15, 2009 at 4:55 am
Hi!
Try this, without cursors.
Regards,
Liliana.
SELECT TOP 100 n = IDENTITY(INT, 1, 1) INTO #Numbers
FROM
sysobjects a1
CROSS JOIN
sysobjects a2
CROSS JOIN
sysobjects a3
DECLARE @tblVariable table (Roles varchar(2000))
insert into @tblVariable VALUES('Role1,Role2,')
insert into @tblVariable VALUES('Role1,Role3,')
insert into @tblVariable VALUES('Role1,Role4,')
insert into @tblVariable VALUES('Role2,Role5,')
insert into @tblVariable VALUES('Role6,Role1,')
insert into @tblVariable VALUES('Role7,')
insert into @tblVariable VALUES('Role8,')
DECLARE @tblRolesInSingleColumn TABLE(RoleId VARCHAR(8000))
INSERT @tblRolesInSingleColumn
SELECT DISTINCT
SUBSTRING (t.Roles + ',', N.n, CHARINDEX (',', t.Roles + ',', N.n) - N.n) as RoleId
FROM @tblVariable t
CROSS JOIN #Numbers N
WHERE SUBSTRING (t.Roles, N.n - 1, 1) = ',' OR N.n = 1
ORDER BY 1
DECLARE @Result varchar(8000)
SET @Result = ''
SELECT @Result = @Result + CASE WHEN LEN(@Result)>0 THEN ', ' ELSE '' END + RoleId
FROM (SELECT DISTINCT RoleId FROM @tblRolesInSingleColumn)Roles
SELECT @Result RolesIds
GO
January 15, 2009 at 9:56 pm
it is great, Thank you Liliana.
January 21, 2009 at 8:42 am
Hi All,
In the above code, i have noticed a problem that in the #Numbers table,
It is not a problem, just suppose if the query results more than the no.of the rows in the identity column, then i will not give.
In this scenario, we have inserted 100 numbers in the #numbers, so it will return up to 100.
...
one person suggested the other logic... in the same fourm
CREATE TABLE #t2(ID INT IDENTITY, Column1 NVARCHAR(100))
CREATE TABLE #t3(Column1 NVARCHAR(100))
INSERT INTO #t2
SELECT 'Value1, Value2'
UNION ALL SELECT 'Value1, Value3'
UNION ALL SELECT 'Value1, Value2'
UNION ALL SELECT 'Value4, Value5'
UNION ALL SELECT 'Value1, Value3'
UNION ALL SELECT 'Value1, Value2'
UNION ALL SELECT 'Value4, Value5'
SELECT * FROM #t2
UPDATE #t2 SET Column1 = replace(Column1,' ','')+','
SELECT * FROM #t2
DECLARE @Values NVARCHAR(100), @ID INT
SELECT @ID = MIN(ID) FROM #t2
WHILE @ID is NOT NULL
BEGIN
SELECT @Values = SUBSTRING(Column1, 1,PATINDEX('%,%',Column1)) FROM #t2 WHERE ID = @ID
WHILE ISNULL(@Values,'') <> ''
BEGIN
INSERT INTO #t3
SELECT @Values
UPDATE #t2 SET Column1 = STUFF(Column1, 1, LEN(@Values), '') WHERE ID = @ID
SELECT @Values = SUBSTRING(Column1, 1,PATINDEX('%,%',Column1)) FROM #t2 WHERE ID = @ID
END
SELECT @ID = MIN(ID) FROM #t2 WHERE ID > @ID
END
UPDATE #t3 SET Column1 = replace(Column1,',','')
SELECT DISTINCT *
FROM #t3
DROP TABLE #t2
DROP TABLE #t3
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply