T-SQL results in to columns - sql

Question

I have a table (t1) like below
Id Name RelId
1 a 2
2 b 3
3 c 4
4 d 3
5 e 6
The other table (t2)
Id data FK Order
1 aa 2 2
2 bb 2 3
3 cc 2 1
4 dd 2 4
5 ee 2 5
6 ff 3 3
7 gg 3 2
8 hh 3 1
9 ii 4 7
10 jj 4 4
11 kk 4 1
12 ll 4 3
13 mm 6 1
14 nn 6 2
15 oo 6 3
16 pp 6 4
My output result am looking for is
+----+------+-------+-------+------+----------+
| id | name | RelId | Col 1 | Col2 | Col-Oth |
+----+------+-------+-------+------+----------+
| 1 | a | 2 | cc | aa | bb,dd,ee |
| 2 | b | 3 | hh | gg | ff |
| 3 | c | 4 | kk | ll | jj,ii |
| 4 | d | 3 | hh | gg | ff |
| 5 | e | 6 | mm | nn | oo,pp |
+----+------+-------+-------+------+----------+
based on the Relid in T1 table join with FK column in T2 and populate col1 with the least order data, col2 with the next higher order data and col-oth with remaining data comma separated ordered.
Need your help on same.
SELECT id,name,RelId, (select data,rownumber() (partition by data order by order asc) from t2 inner join t1 on t1.relid= t2.FK) from t1

Answer 1

Try following query:
DECLARE #TEMP TABLE
(
Id INT,
Name VARCHAR(10),
RelId INT
)
INSERT INTO #TEMP VALUES (1,'a',2),(2,'b',3),(3,'c',4),(4,'d',3),(5,'e',6)
DECLARE #TEMP1 TABLE
(
Id INT,
Data varchar(10),
FK INT,
[order] INT
)
INSERT INTO #TEMP1 VALUES
(1 ,'aa',2,2),(2 ,'bb',2,3),(3 ,'cc',2,1),(4 ,'dd',2,4),(5 ,'ee',2,5),
(6 ,'ff',3,3),(7 ,'gg',3,2),(8 ,'hh',3,1),(9 ,'ii',4,7),(10,'jj',4,4),
(11,'kk',4,1),(12,'ll',4,3),(13,'mm',6,1),(14,'nn',6,2),(15,'oo',6,3),(16,'pp',6,4)
SELECT
t1.*,
(SELECT Data FROM (SELECT ROW_NUMBER() OVER(ORDER BY t2.[order]) As RowNo,Data FROM #TEMP1 t2 WHERE t2.FK = t1.RelId)t3 WHERE t3.RowNo=1),
(SELECT Data FROM (SELECT ROW_NUMBER() OVER(ORDER BY t2.[order]) As RowNo,Data FROM #TEMP1 t2 WHERE t2.FK = t1.RelId)t3 WHERE t3.RowNo=2),
STUFF((SELECT DISTINCT ',' + Data FROM (SELECT ROW_NUMBER() OVER(ORDER BY t2.[order]) As RowNo,Data FROM #TEMP1 t2 WHERE t2.FK = t1.RelId)t3 WHERE t3.RowNo > 2 FOR XML PATH ('')), 1, 1, '')
FROM
#TEMP t1

Answer 2

Using PIVOT:
DECLARE #t1 TABLE
(
ID INT ,
Name CHAR(1) ,
RelID INT
)
DECLARE #t2 TABLE
(
ID INT ,
Data CHAR(2) ,
RelID INT ,
Ordering INT
)
INSERT INTO #t1
VALUES ( 1, 'a', 2 ),
( 2, 'b', 3 ),
( 3, 'c', 4 ),
( 4, 'd', 3 ),
( 5, 'e', 6 )
INSERT INTO #t2
VALUES ( 1, 'aa', 2, 2 ),
( 2, 'bb', 2, 3 ),
( 3, 'cc', 2, 1 ),
( 4, 'dd', 2, 4 ),
( 5, 'ee', 2, 5 ),
( 6, 'ff', 3, 3 ),
( 7, 'gg', 3, 2 ),
( 8, 'hh', 3, 1 ),
( 9, 'ii', 4, 7 ),
( 10, 'jj', 4, 4 ),
( 11, 'kk', 4, 1 ),
( 12, 'll', 4, 3 ),
( 13, 'mm', 6, 1 ),
( 14, 'nn', 6, 2 ),
( 15, 'oo', 6, 3 ),
( 16, 'pp', 6, 4 );
WITH cte1
AS ( SELECT t1.ID ,
t1.Name ,
t1.RelID ,
t2.Data ,
ROW_NUMBER() OVER ( PARTITION BY t1.ID ORDER BY t2.Ordering ) AS rn
FROM #t1 t1
JOIN #t2 t2 ON t1.RelID = t2.RelID
),
cte2
AS ( SELECT ID ,
Name ,
RelID ,
Data ,
rn ,
STUFF(( SELECT ',' + Data
FROM cte1 ci
WHERE co.ID = ci.ID
AND rn > 2
FOR
XML PATH('')
), 1, 1, '') AS Col3
FROM cte1 co
)
SELECT ID ,
Name ,
RelID ,
[1] AS Col1 ,
[2] AS Col2 ,
Col3
FROM cte2 PIVOT( MAX(data) FOR rn IN ( [1], [2] ) ) p
Output:
ID Name RelID Col1 Col2 Col3
1 a 2 cc aa bb,dd,ee
2 b 3 hh gg ff
3 c 4 kk ll jj,ii
4 d 3 hh gg ff
5 e 6 mm nn oo,pp
Execution plan of my statement
Execution plan of accepted statement:
Which is better? :)