Produce result table trom multiple tables - sql

Question

SQL Server 2008 R2
I have 3 tables contained data for 3 different types of events
Type1, Type2, Type3 with two columns:
DatePoint ValuePoint
I want to produce result table which would look like that:
DatePoint TotalType1 TotalType2 TotalType3
I've started from that
SELECT [DatePoint]
,SUM(ValuePoint) as TotalType1
FROM [dbo].[Type1]
GROUP BY [DatePoint]
ORDER BY [DatePoint]
SELECT [DatePoint]
,SUM(ValuePoint) as TotalType2
FROM [dbo].[Type2]
GROUP BY [DatePoint]
ORDER BY [DatePoint]
SELECT [DatePoint]
,SUM(ValuePoint) as TotalType3
FROM [dbo].[Type3]
GROUP BY [DatePoint]
ORDER BY [DatePoint]
So I have three result but I need to produce one (Date TotalType1 TotalType2 TotalType3), what I need to do next achieve my goal?
UPDATE
Forgot to mention that DatePoint which is exists in one type may or may not exist in another

Answer 1

Here's my take. I assume that you don't have the same datetime values in every table (certainly, the stuff I get to work with is never so consistant). There should be an easier way to do this, but once you're past two outer joins things can get pretty tricky.
SELECT
dp.DatePoint
,isnull(t1.TotalType1, 0) TotalType1
,isnull(t2.TotalType2, 0) TotalType2
,isnull(t3.TotalType3, 0) TotalType3
from (-- Without "ALL", UNION will filter out duplicates
select DatePoint
from Type1
union select DatePoint
from Type2
union select DatePoint
from Type3) dp
left outer join (select DatePoint, sum(ValuePoint) TotalType1
from Type1
group by DatePoint) t1
on t1.DatePoint = db.DatePoint
left outer join (select DatePoint, sum(ValuePoint) TotalType2
from Type2
group by DatePoint) t2
on t2.DatePoint = db.DatePoint
left outer join (select DatePoint, sum(ValuePoint) TotalType3
from Type3
group by DatePoint) t3
on t3.DatePoint = db.DatePoint
order by dp.DatePoint

Answer 2

Suppose some distinct could help, but the general idea should be the following:
SELECT
t.[DatePoint],
SUM(t1.ValuePoint) as TotalType1,
SUM(t2.ValuePoint) as TotalType2,
SUM(t3.ValuePoint) as TotalType3
FROM
(
SELECT [DatePoint] FROM [dbo].[Type1]
UNION
SELECT [DatePoint] FROM [dbo].[Type2]
UNION
SELECT [DatePoint] FROM [dbo].[Type3]
) as t
LEFT JOIN
[dbo].[Type1] t1
ON
t1.[DatePoint] = t.[DatePoint]
LEFT JOIN
[dbo].[Type2] t2
ON
t2.[DatePoint] = t.[DatePoint]
LEFT JOIN
[dbo].[Type3] t3
ON
t3.[DatePoint] = t.[DatePoint]
GROUP BY
t.[DatePoint]
ORDER BY
t.[DatePoint]

Answer 3

To avoid all of the JOINs:
SELECT
SQ.DatePoint,
SUM(CASE WHEN SQ.type = 1 THEN SQ.ValuePoint ELSE 0 END) AS TotalType1,
SUM(CASE WHEN SQ.type = 2 THEN SQ.ValuePoint ELSE 0 END) AS TotalType2,
SUM(CASE WHEN SQ.type = 3 THEN SQ.ValuePoint ELSE 0 END) AS TotalType3
FROM (
SELECT
1 AS type,
DatePoint,
ValuePoint
FROM
dbo.Type1
UNION ALL
SELECT
2 AS type,
DatePoint,
ValuePoint
FROM
dbo.Type2
UNION ALL
SELECT
3 AS type,
DatePoint,
ValuePoint
FROM
dbo.Type3
) AS SQ
GROUP BY
DatePoint
ORDER BY
DatePoint
From the little information provided though, it seems like there are some flaws in the database design, which is probably part of the reason that querying the data is so difficult.