SQL Server aggregate over range of dates - sql

Question

I am using SQL Server 2014. I need to aggregate totals (sum total) over a range of dates that are partitioned or grouped by customer and location. The key is to get all the adjustment amounts and sum them up as they apply to a billing transaction date.
So all adjustments after the last bill date, but less than the next bill date need to get summed up and presented nicely along with the bill amount.
See example:
+------------------+------------+------------+------------------+--------------------+
| TRANSACTION_TYPE | CUSTOMERID | LOCATIONID | TRANSACTION DATE | TRANSACTION AMOUNT |
+------------------+------------+------------+------------------+--------------------+
| bill | 215 | 102 | 7/7/2016 | $100.00 |
| bill | 215 | 102 | 6/6/2016 | $121.00 |
| adj | 215 | 102 | 6/1/2016 | $22.00 |
| adj | 215 | 102 | 5/8/2016 | $0.35 |
| adj | 215 | 102 | 5/7/2016 | $5.00 |
| bill | 215 | 102 | 5/6/2016 | $115.00 |
| bill | 215 | 102 | 4/7/2016 | $200.00 |
| adj | 215 | 102 | 4/2/2016 | $4.35 |
| adj | 215 | 102 | 4/1/2016 | $(0.50) |
| adj | 215 | 102 | 3/28/2016 | $33.00 |
| bill | 215 | 102 | 3/28/2016 | $75.00 |
| adj | 215 | 102 | 3/5/2016 | $0.33 |
| bill | 215 | 102 | 3/3/2016 | $99.00 |
+------------------+------------+------------+------------------+--------------------+
What I would like to see is the following:
+------------------+------------+------------+------------------+-------------+-------------------+
| TRANSACTION_TYPE | CUSTOMERID | LOCATIONID | TRANSACTION DATE | BILL AMOUNT | ADJUSTMENT AMOUNT |
+------------------+------------+------------+------------------+-------------+-------------------+
| bill | 215 | 102 | 7/7/2016 | $100.00 | $- |
| bill | 215 | 102 | 6/6/2016 | $121.00 | $27.35 |
| bill | 215 | 102 | 5/6/2016 | $115.00 | $- |
| bill | 215 | 102 | 4/7/2016 | $200.00 | $36.85 |
| bill | 215 | 102 | 3/28/2016 | $75.00 | $0.33 |
| bill | 215 | 102 | 3/3/2016 | $99.00 | $- |
+------------------+------------+------------+------------------+-------------+-------------------+

Answer 1

You need to:
first conceive the table as two (virtual) sub-tables, on the TransactionType;
then use the LEAD function to get the date range of adjustments to be applied; and
finally perform a eft join.
Untested SQL below:
with
BillData as (
select
TransactionType,
CustomerID,
LocationID,
TransactionDate,
TransactionAmount,
lead(TransactionDate, 1) over (partition by CustomerID
order by TransactionDate) as NextDate
from #data bill
where TransactionType = 'bill'
),
AdjData as (
select
CustomerID,
TransactionDate,
sum(TransactionAmount) as AdjAmount
from #data adj
where TransactionType = 'adj'
)
select
bill.TransactionType,
bill.CustomerID,
bill.LocationID,
bill.TransactionDate,
sum(TransactionAmount) as BillAmount,
sum(AdjAmount) as AdjAmount
from BillData bill
left join AdjData adj
on adj.CustomerID = bill.CustomerID
and bill.TransactionDate <= adj.TransactionDate
and adj.TransactionDate < bill.NextDate
group by
bill.TransactionType,
bill.CustomerID,
bill.LocationID,
bill.TransactionDate
;

Answer 2

This is what I ended up doing:
select
bill.TransactionType,
bill.CustomerID,
bill.LocationID,
bill.TransactionDate,
TransactionAmount as BillAmount,
sum(AdjAmount) as AdjAmount
from
(
select
TransactionType,
CustomerID,
LocationID,
TransactionDate,
TransactionAmount,
lag(TransactionDate, 1) over (partition by CustomerID, LocationID
order by TransactionDate) as PreviousDate --NextDate
from test1
where TransactionType = 'bill'
) as bill
left join
(
select
CustomerID,
LocationID,
TransactionDate,
TransactionAmount as AdjAmount
from test1
where TransactionType = 'adj'
) as adj
ON
adj.CustomerID = bill.CustomerID
and adj.LocationID = bill.LocationID
and adj.TransactionDate >= bill.PreviousDate
and adj.TransactionDate < bill.TransactionDate
group by
bill.TransactionType,
bill.CustomerID,
bill.LocationID,
bill.TransactionDate,
bill.TransactionAmount
order by 4 desc