repeatation on output result according date insqlserver - sql-server

Question

Concatenate values from multiple columns of 3 tables and display in one column. Date field is datetime.now my problem is if date is repeat of one table then amt2,amt3 is also repeat on same date
table1:
date amt1
1-1-2016 10:44:00 111
1-1-2016 10:44:02 222 repeat date
3-4-2016 11:44:00 111
table2:
date amt2
1-1-2016 10:44:02 101
2-2-2016 10:44:02 333
2-3-2016 10:44:02 444
table3:
date amt3
2-2-2016 10:44:04 001
2-3-2016 10:44:02 002
3-3-2016 10:44:02 003
query
SELECT
COALESCE(t1.date, t2.date, t3.date)
, t1.amt1
, t2.amt2
, t3.amt3
FROM tb_amount t1
FULL JOIN tb_cashamt t2 ON t1.date =t2.date
FULL JOIN tb_grandtotal t3 ON t2.date = t3.date
output:
date amt1 amt2 amt3
1-1-2016 10:44:00 111 null NULL
1-1-2016 10:44:02 111 101 null data repeat on amt1(111)
2-2-2016 10:44:02 NULL 333 null
2-2-2016 10:44:04 NULL 333 001 data repeat on amt2(111)
TOTAL --- --- -----
Desired output:
date amt1 amt2 amt3
1-1-2016 111 101 NULL
2-2-2016 NULL 333 001
TOTAL total(amt1) total(amt1) total(amt1)
I have problem with smalldatetime datatype on datecolumn. I want a result on each date and amt1,amt2,amt3 show total data on this date

Answer 1

Use FULL JOIN
SELECT
COALESCE(t1.date, t2.date, t3.date)
, t1.amt1
, t2.amt2
, t3.amt3
FROM Table1 t1
FULL JOIN Table2 t2 ON t1.date =t2.date
FULL JOIN Table3 t3 ON t2.date = t3.date

Answer 2

You'll need to use the FULL JOIN, also known as FULL OUTER JOIN as explained on Using Outer Joins:
To retain the nonmatching information by including nonmatching rows in the results of a join, use a full outer join. SQL Server provides the full outer join operator, FULL OUTER JOIN, which includes all rows from both tables, regardless of whether or not the other table has a matching value.
;
WITH Table1 AS
( SELECT [Date], Amt1
FROM ( VALUES
('2016-01-01', 111),
('2016-02-02', 222),
('2016-04-03', 111)
) AS Table1(Date, Amt1)
), Table2 AS
( SELECT [Date], Amt2
FROM ( VALUES
('2016-01-01', 101),
('2016-02-02', 333),
('2016-03-02', 444),
('2016-03-03', 456),
('2016-04-01', 101),
('2016-04-03', 111)
) AS Table2(Date, Amt2)
), Table3 AS
( SELECT [Date], Amt3
FROM ( VALUES
('2016-02-02', 001),
('2016-03-02', 002),
('2016-03-03', 003),
('2016-04-01', 555),
('2016-04-02', 666),
('2016-04-03', 777)
) AS Table3(Date, Amt3)
)
SELECT ISNULL(Table1.Date, ISNULL(Table2.Date, Table3.Date)) Date,
Table1.Amt1, Table2.Amt2, Table3.Amt3
FROM Table1
FULL JOIN Table2 ON Table1.Date = Table2.Date
FULL JOIN Table3 ON Table2.Date = Table3.Date
ORDER BY ISNULL(Table1.Date, ISNULL(Table2.Date, Table3.Date))

Answer 3

If you need only one row per date, you need to group by + sum each of the tables on date level first before joining + you can use rollup for the total:
select coalesce(T1.Date,T2.Date,T3.Date), sum(T1.Amt1),Sum(T2.Amt2),Sum(T3.Amt3)
from
(
select convert(date, date) as Date, sum(Amt1) as Amt1
from Table1
group by convert(date, date)
) T1
full outer join (
select convert(date, date) as Date, sum(Amt2) as Amt2
from Table2
group by convert(date, date)
) T2 on T1.Date = T2.Date
full outer join (
select convert(date, date) as Date, sum(Amt3) as Amt3
from Table3
group by convert(date, date)
) T3 on isnull(T1.Date,T2.Date) = T3.Date
group by rollup(coalesce(T1.Date,T2.Date,T3.Date))