I have three tables:
Product
ProductID ProductName
1 Cycle
2 Scooter
3 Car
<
You can use SQL Server's PIVOT operator
SELECT *
FROM (
SELECT P.ProductName
, C.CustName
, T.Amount
FROM Transactions AS T
INNER JOIN Product AS P ON T.ProductID = P.ProductID
INNER JOIN Customer AS C ON T.CustomerID = C.CustomerID
WHERE T.TranDate BETWEEN '2011-01-01' AND '2011-03-31'
) s
PIVOT (SUM(Amount) FOR ProductName IN ([Car], [Cycle], [Scooter])) pvt
Test data
;WITH q AS (
SELECT [Product] = 'Car', [Customer] = 'Armstrong', [Amount] = 80115.50
UNION ALL SELECT 'Car', 'Michelle', 36571.85
UNION ALL SELECT 'Car', 'Schmidt', 45000.65
UNION ALL SELECT 'Cycle', 'Michelle', 15000.00
UNION ALL SELECT 'Cycle', 'Ronald', 25000.00
UNION ALL SELECT 'Scooter', 'Peterson', 82658.23
UNION ALL SELECT 'Scooter', 'Ronald', 98547.52
UNION ALL SELECT 'Scooter', 'Schmidt', 54000.25
)
SELECT Customer
, Car = ISNULL(Car, 0)
, Cycle = ISNULL(Cycle, 0)
, Scooter = ISNULL(Scooter, 0)
, Total = ISNULL(Car, 0) + ISNULL(Cycle, 0) + ISNULL(Scooter, 0)
FROM (
SELECT *
FROM q
) s
PIVOT (SUM(Amount) FOR Product IN ([Car], [Cycle], [Scooter])) pvt
Output
Customer Car Cycle Scooter Total
Armstrong 80115.50 0.00 0.00 80115.50
Michelle 36571.85 15000.00 0.00 51571.85
Peterson 0.00 0.00 82658.23 82658.23
Ronald 0.00 25000.00 98547.52 123547.52
Schmidt 45000.65 0.00 54000.25 99000.90
create table #Product (ProductID int,ProductName varchar(15))
insert into #Product values (1,'Cycle')
insert into #Product values (2,'Scooter')
insert into #Product values (3,'Car')
create table #Customer (CustomerID int, CustomerName varchar(30))
insert into #Customer values (101,'Ronald')
insert into #Customer values (102,'Michelle')
insert into #Customer values (103,'Armstrong')
insert into #Customer values (104,'Schmidt')
insert into #Customer values (105,'Peterson')
create table #Transactions (TID int,ProductID int,CustomerID int, TranDate smalldatetime,Amount decimal(18,2))
insert into #Transactions values (10001,1,101,'01-Jan-11',25000.00)
insert into #Transactions values (10002,2,101,'02-Jan-11',98547.52)
insert into #Transactions values (10003,1,102,'03-Feb-11',15000.00)
insert into #Transactions values (10004,3,102,'07-Jan-11',36571.85)
insert into #Transactions values (10005,2,105,'09-Feb-11',82658.23)
insert into #Transactions values (10006,2,104,'10-Feb-11',54000.25)
insert into #Transactions values (10007,3,103,'20-Feb-11',80115.50)
insert into #Transactions values (10008,3,104,'22-Feb-11',45000.65)
with temp as
(
select cus.CustomerName,pro.ProductName, sum(trans.Amount) as Amount from #Transactions as trans
inner join #Customer as cus on trans.CustomerID = cus.CustomerID
inner join #Product as pro on trans.ProductID = pro.ProductID
group by cus.CustomerName,pro.ProductName
)
select CustomerName,isnull([Car],0)Car, isnull([Cycle],0)Cycle,isnull([Scooter],0) as Scooter, isnull([Car],0)+isnull([Cycle],0)+isnull([Scooter],0)as Total from temp
pivot (
sum(Amount) for ProductName in ([Cycle],[Scooter],[Car])
)pot*
We can create matrix using pivot, this can easily done with data frames
product|Key|Value
A |P |10|
A |Q |40|
B |R |50|
B |S |50|
val newdf=df.groupBy("product").pivot("key").sum("value")
|product|P |Q |R |S |
|B |null|null| 50| 50|
|A | 10| 40|null|null|
We can replace null and we can do calculations as well