SQL Server 2005. I am not after a coded answer here (although it would be nice). I'm really after advice on the best way forward to get the result I need. I have some knowledge of pivot/unpivot/cte//rownumber and dynamic queries but cannot get my head around this particular problem! An example of the data follows. Note: The occurrence of type,location,name and description can be none to many.
drop table #temp
create table #temp
(
event int,
type varchar(20),
locations varchar(20),
name varchar(30),
description varchar(50)
)
insert into #temp values (1,'support','r1','fred','desc 1')
insert into #temp values (1,'support','r1','fred','desc 2')
insert into #temp values (1,'support','r1','fred','desc 3')
insert into #temp values (1,'support','r1','jim','desc 1')
insert into #temp values (1,'support','r1','jim','desc 2')
insert into #temp values (1,'support','r1','jim','desc 3')
insert into #temp values (1,'support','r2','fred','desc 1')
insert into #temp values (1,'support','r2','fred','desc 2')
insert into #temp values (1,'support','r2','fred','desc 3')
insert into #temp values (1,'support','r2','jim','desc 1')
insert into #temp values (1,'support','r2','jim','desc 2')
insert into #temp values (1,'support','r2','jim','desc 3')
insert into #temp values (1,'work','r1','fred','desc 1')
insert into #temp values (1,'work','r1','fred','desc 2')
insert into #temp values (1,'work','r1','fred','desc 3')
insert into #temp values (1,'work','r1','jim','desc 1')
insert into #temp values (1,'work','r1','jim','desc 2')
insert into #temp values (1,'work','r1','jim','desc 3')
insert into #temp values (1,'work','r2','fred','desc 1')
insert into #temp values (1,'work','r2','fred','desc 2')
insert into #temp values (1,'work','r2','fred','desc 3')
insert into #temp values (1,'work','r2','jim','desc 1')
insert into #temp values (1,'work','r2','jim','desc 2')
insert into #temp values (1,'work','r2','jim','desc 3')
select * from #temp
The result I am after is this ..
1,support;work,r1;r2,fred;jim,desc1;desc2;desc3
Your goal seem to select all distinct value of all columns, then Concatenate into one string. And you only need advice, so I recommend you go here: multiple rows into a single row
It seem that you need more help:
select distinct
stuff((SELECT distinct'; ' + type-- as type
FROM #temp
--order by type
FOR XML PATH('')),1,1,'')
+ (SELECT distinct'; ' + locations
FROM #temp
FOR XML PATH(''))
+ (SELECT distinct'; ' + name
FROM #temp
FOR XML PATH(''))
+ (SELECT distinct'; ' + description
FROM #temp
FOR XML PATH(''))
from #temp;
If you need 4 columns, then change + (SELECT to , stuff((SELECT
The query is just that simple: get distinct of one column, change into string, then concatenate + string of (next column)...
This is slightly unrelated, but when inserting data like this it would be easier (for you) to do it like this (also, try to get into the habit of naming the fields you are inserting into);
INSERT INTO #temp (event, type, locations, name, description)
VALUES (1,'support','r1','fred','desc 1')
,(1,'support','r1','fred','desc 2')
,(1,'support','r1','fred','desc 3')
,(1,'support','r1','jim','desc 1')
,(1,'support','r1','jim','desc 2')
Please don't up-vote this version! All up-votes should go to the solution above! The code below is simply for completeness. This shows the syntax for organising the data into separate columns based on the answer from #NayruLove
SELECT distinct x.event,
stuff((SELECT distinct'; ' + t.type
FROM #temp t
where t.event = x.event
FOR XML PATH('')),1,1,'') as type
, stuff((SELECT distinct'; ' + locations
FROM #temp t
where t.event= x.event
FOR XML PATH('')),1,1,'') as room
, stuff((SELECT distinct'; ' + name
FROM #temp t
where t.event = x.event
FOR XML PATH('')),1,1,'') as name
, stuff((SELECT distinct'; ' + description
FROM #temp t
where t.event = x.event
FOR XML PATH('')),1,1,'') as description
from #temp x
group by x.event
Related
I need to display a query output in a horizontal manner. I have some example data
create table TestTable (id number, name varchar2(10))
insert into TestTable values (1, 'John')
insert into TestTable values (2, 'Mckensy')
insert into TestTable values (3, 'Valneech')
insert into TestTable values (4, 'Zeebra')
select * from TestTable
This gets the output in a vertical view.
ID Name
==========
1 John
2 Mckensy
3 Valneech
4 Zeebra
However, I need to display it horizontally.
ID 1 2 3 4
Name John Mckensy Valneech Zeebra
create table #TestTable (id int, name varchar(10))
insert into #TestTable values (1, 'John')
insert into #TestTable values (2, 'Mckensy')
insert into #TestTable values (3, 'Valneech')
insert into #TestTable values (4, 'Zeebra')
select *
from
(
select *
from #TestTable
) src
pivot
(
max(name)
for id in ([1], [2], [3],[4])
) piv;
output
1 2 3 4
John Mckensy Valneech Zeebra
You can also use dynamic sql query something like below.
Query
declare #sql as varchar(max);
select #sql = 'select ' + char(39) + 'Name' + char(39) + ' Id,' + stuff((
select
',max(case [Id] when ' + cast(id as varchar(10)) + ' then name end) ['
+ cast([Id] as varchar(10)) + ']'
from TestTable
for xml path('')
), 1, 1, '');
select #sql += 'from TestTable;';
exec(#sql);
my Column is like this:
col1
Mary
Jack
John
and i need to get something like this:
Mary,Jack,John
I did it by using "For XML PATH", but the problem is that segment not work in SQL server view,
is there other way to get concatenation of one column?
try using SQL COALESCE experession to convert your column values into a comma separated string.
here's what you need to do.
DECLARE #test TABLE (
SampleCol varchar(50)
)
INSERT INTO #Test
VALUES ('Test')
INSERT INTO #Test
VALUES ('Test 1')
INSERT INTO #Test
VALUES ('Test 2')
INSERT INTO #Test
VALUES ('Test 3')
INSERT INTO #Test
VALUES ('Test 4')
DECLARE #aa varchar(200)
SET #aa = ''
SELECT
#aa =
COALESCE(CASE
WHEN #aa = '' THEN SampleCol
ELSE #aa + ',' + SampleCol
END
, '')
FROM #test
SELECT
#aa
here's a complete SQLFIDDLE
Try this:
DECLARE #tbl TABLE(col1 VARCHAR(10));
INSERT INTO #tbl VALUES('Mary'),('Jack'),('John');
SELECT
STUFF
(
(
SELECT ', ' + col1
FROM #tbl
FOR XML PATH('')
)
,1,2,'')
I have a table variable #Holding two columns: an id (not unique) and a message:
id message
---- -------
2 give
2 me
2 help
3 Need
3 help
1 help!
The result should be
2 give me help
3 Need help
1 help!
This it very much simplified, but shows that there are id which may exist more than once, and some kind of text which should be concatenated into a string.
I cannot manage it to loop through this table variable (but not through a table too!).
I tried a cursor (which I did not understand correctly) but it failed of course.
The number of records are not that much, not even 100 in that table variable.
Thanks yr. help
Michael
Original question
No CURSOR, WHILE loop, or User-Defined Function needed.
Just need to be creative with FOR XML and PATH.
[Note: This solution only works on SQL 2005 and later. Original question didn't specify the version in use.]
CREATE TABLE #YourTable ([ID] INT, [Name] CHAR(1), [Value] INT)
INSERT INTO #YourTable ([ID],[Name],[Value]) VALUES (1,'A',4)
INSERT INTO #YourTable ([ID],[Name],[Value]) VALUES (1,'B',8)
INSERT INTO #YourTable ([ID],[Name],[Value]) VALUES (2,'C',9)
SELECT
[ID],
STUFF((
SELECT ', ' + [Name] + ':' + CAST([Value] AS VARCHAR(MAX))
FROM #YourTable
WHERE (ID = Results.ID)
FOR XML PATH(''),TYPE).value('(./text())[1]','VARCHAR(MAX)')
,1,2,'') AS NameValues
FROM #YourTable Results
GROUP BY ID
DROP TABLE #YourTable
This should work if you are using SQL Server.
select T1.id,
stuff((select ' '+T2.[message]
from #A as T2
where T1.id = T2.id
for xml path(''), type).value('.', 'varchar(max)'), 1, 1, '') as [message]
from #A as T1
group by T1.id
Pretty the same, just less letters in the code:
DECLARE #t AS TABLE ( id INT, msg VARCHAR(100) );
INSERT INTO #t
VALUES ( 2, 'give' );
INSERT INTO #t
VALUES ( 2, 'me' );
INSERT INTO #t
VALUES ( 2, 'help' );
INSERT INTO #t
VALUES ( 3, 'Need' );
INSERT INTO #t
VALUES ( 3, 'help' );
INSERT INTO #t
VALUES ( 1, 'help!' );
SELECT DISTINCT
id ,
(
SELECT ST1.msg + ' '
FROM #t ST1
WHERE ST1.id = ST2.id
FOR XML PATH('')
) t
FROM
#t ST2;
If I have a team table with a unknown amount of members, is there a way to make the pivot query dynamic?
create table #t (
team varchar (20), member varchar (20)
)
insert into #t values ('ERP', 'Jack')
insert into #t values ('ERP', 'John')
insert into #t values ('ERP', 'Mary')
insert into #t values ('ERP', 'Tim')
insert into #t values ('CRM', 'Robert')
insert into #t values ('CRM', 'Diana')
select * from #t
select team, [1] as teamMember1, /* 1st select */
[2] as teamMember2, [3] as teamMember3
from
(select team , member, row_number () /* 3rd select */
over (partition by team order by team) as rownum
from #t) a
pivot (max(member) for rownum in ([1], [2], [3])) as pvt
drop table #t
Why yes, yes there is. Here's a script I cooked up years ago for a similar problem that was ultimately solved by giving the user Excel and washing my hands of it. I apologize it's not configured with your example data, but hopefully it's easy to follow.
Hope that helps,
John
--------------START QUERY--------------
-- Example Table
CREATE TABLE #glbTestTable
(
ProviderID INT,
Total INT,
PaymentDate SMALLDATETIME
)
--So the dates insert properly
SET DATEFORMAT dmy
-- Populate Example Table
INSERT INTO #glbTestTable VALUES (232, 12200, '12/01/09')
INSERT INTO #glbTestTable VALUES (456, 10200, '12/01/09')
INSERT INTO #glbTestTable VALUES (563, 11899, '02/03/09')
INSERT INTO #glbTestTable VALUES (221, 5239, '13/04/09')
INSERT INTO #glbTestTable VALUES (987, 7899, '02/03/09')
INSERT INTO #glbTestTable VALUES (1, 1234, '02/08/09')
INSERT INTO #glbTestTable VALUES (2, 4321, '02/07/09')
INSERT INTO #glbTestTable VALUES (3, 5555, '02/06/09')
-- Raw Output
SELECT *
FROM #glbTestTable
-- Build Query for Pivot --
DECLARE #pvtColumns VARCHAR(MAX)
SET #pvtColumns = ''
-- Grab up to the first 1023 "Columns" that we want to use in Pivot Table.
-- Tables can only have 1024 columns at a maximum
SELECT TOP 1023 #pvtColumns = #pvtColumns + '[' + CONVERT(VARCHAR, PaymentDate, 103) + '], '
FROM (SELECT DISTINCT PaymentDate FROM #glbTestTable) t_distFP
-- Create PivotTable Query
DECLARE #myQuery VARCHAR(MAX)
SET #myQuery = '
SELECT ProviderID, ' + LEFT(#pvtColumns, LEN(#pvtColumns) - 1) + '
FROM (SELECT ProviderID, PaymentDate, Total
FROM #glbTestTable) AS SourceTable
PIVOT
(
SUM(Total)
FOR PaymentDate IN (' + LEFT(#pvtColumns, LEN(#pvtColumns) - 1) + ')
) AS PivotTable'
-- Run the Pivot Query
EXEC(#myQuery)
-- Cleanup
DROP TABLE #glbTestTable
---------------END QUERY---------------
I have a simple problem , Although i believe its simple , am not able to figure out the same.
Consider i have the below table with exactly same data as given below :
CREATE TABLE #temp
(
link varchar(255),
number INT,
fname varchar(255)
)
insert into #temp VALUES ('abc',1,'f1')
insert into #temp VALUES ('abc',2,'f2')
insert into #temp VALUES ('abc',3,'f3')
insert into #temp VALUES ('abc',4,'f6')
insert into #temp VALUES ('abc',10,'f100')
insert into #temp VALUES ('abe',-1,'f0')
insert into #temp VALUES ('abe',1,'f1')
insert into #temp VALUES ('abe',2,'f2')
insert into #temp VALUES ('abe',3,'f3')
insert into #temp VALUES ('abe',4,'f6')
insert into #temp VALUES ('abe',20,'f200')
insert into #temp VALUES ('cbe',-1,'f0')
insert into #temp VALUES ('cbe',1,'f1')
insert into #temp VALUES ('cbe',2,'f2')
insert into #temp VALUES ('cbe',3,'f3')
Now for a given link , i need to get the max 'number' and the corresponding 'fname' which has the max 'number' for the given 'link'.
1)Ex : if link is 'abc' , output should be
abc, 10, f100
2)Ex : if link if 'abe' , Output should be
abe, 20, f200
3)Now link can be also given as a pattern , like (link like 'ab%') , so output should be
abc, 10, f100
abe, 20, f200
4)if (link like 'cb%') , so output should be
cbe, 3, f3
Any help in writing this group by query. I have a solution using CAST and string concat like below , but that seems to be in-efficient.
select link,number,fname from #temp
where link like 'ab%' and link+'_'+CAST(number AS varchar(255))
in (select link+'_'+CAST(MAX(number) AS varchar(255)) from #temp
group by link)
Thanks..
Using a self join:
SELECT x.link,
x.number,
x.fname
FROM #temp x
JOIN (SELECT t.link,
MAX(t.number) AS max_number
FROM #temp t
GROUP BY t.link) y ON y.link = x.link
AND y.max_number = x.number
Using a CTE and ROW_NUMBER (SQL Server 2005+):
WITH cte AS (
SELECT x.link,
x.number,
x.fname,
ROW_NUMBER() OVER(PARTITION BY x.link
ORDER BY x.number DESC) rank
FROM #temp x)
SELECT c.link,
c.number,
c.fname
FROM cte c
WHERE c.rank = 1