Aggregation Calculation for All Items with Data rolling from Some Previous Data Points - sql

Question

How do I do aggregation calculations on data rolling from previous data points?
For example, I have a table for the property value evaluations. Each data point is only for a property.
"Property Value Evaluation"
Date Property Value
1/5/2017 A 10
2/3/2017 B 8
2/20/2017 B 12
3/1/2017 A 9
4/10/2017 B 15
And assuming that the property value stays the same as it was evaluated last time, until it is evaluated again. For example, the value of Property A was 10 on 1/5, and it stayed 10 until 3/1 when it was re-evaluated as 9.
How can I have a report to show the trend of the values of all the properties? That is, I need to include all the properties on every data point of the report, such as:
"Value Trend of All Properties"
Date Total Average
1/5/2017 10 10
2/3/2017 18 9
2/20/2017 22 11
3/1/2017 21 10.5
4/10/2017 24 12
(Where Total is the sum of the values of Property A and Property B, and Average is the average value of these two.)
The problem I have is, for the sub-query which takes the Date and the Property as the parameters and returns the last Value, the following error occurs, even though my subquery is "select LAST_VALUE([Value]) ..." which only returns 1 value:
"Subquery returned more than 1 value. This is not permitted when the subquery follows =, !=, <, <= , >, >= or when the subquery is used as an expression."

Answer 1

You can use window functions:
select date, sum(val),
avg(sum(val) * 1.0) over (order by date) as average
from t
group by date
order by date;
You can use window functions in conjunction with aggregation functions.

Answer 2

try this:
SELECT [Date],
Value+ISNULL(LAG(Value) OVER(order by (Select null) ),0) Total ,
(Value+ISNULL(LAG(Value) OVER(order by (Select null) ),0))*1.0/2 As Average
from Property_Evaluation

Answer 3

With the hints from the answers posted here by the others, I built the complete SQL statement as below.
SELECT sq1.Date
, sum(sq1.Value) Total
, avg(sq1.Value) Average
from (
select sq2.Date
, sq2.Property
, ISNULL(sq6.Value, (
select sq3.Value
From (
select sq6.Date
, sq6.Property
, isnull(sq6.Value, LAG(sq6.Value) OVER (partition by sq6.Property order by sq6.Date)) Value
from (
select isnull(sq4.Date, sq5.Date) Date
, isnull(sq4.Property, sq5.Property) Property
, sq4.Value Value
from (
select distinct sq2.Date Date
, sq2.Property Property
from [PropertyEvaluation]
) sq5
full outer join (
SELECT [Date] Date
, [Property] Property
, [Value] Value
FROM [PropertyEvaluation]
where [Property] = sq2.Property) sq4
on sq5.Date = sq4.Date
) sq6
) sq3
where sq3.Date = sq2.Date
and sq3.Property = sq2.Property
)
) Value
from (
select sq7.Date Date
, sq8.Property Property
from (
select distinct [Date] Date
from [PropertyEvaluation]) sq7
, (select distinct [Property] Property
from [PropertyEvaluation]) sq8
) sq2
left join (
SELECT [Date] Date
, [Property] Property
, [Value] Value
FROM [PropertyEvaluation]
) sq6
on sq2.Date = sq6.Date and sq2.Property = sq6.Property
) sq1
group by sq1.Date
order by 1
It returns the desired results:
Date Total Average
1/5/2017 10 10
2/3/2017 18 9
2/20/2017 22 11
3/1/2017 21 10.5
4/10/2017 24 12