SQL Server : sum rows from different tables

Question

Table 1:

 item    Purchase price
 ------- -------------- 
 item1    10$
 item2    20$
 item3    30$

Table 2:

 item    Sale Price
 ------- --------------
 item1    10$
 item2    20$
 item3    30$

I want the following result:

 row name         item1   item2   item3   total
 --------------  ------  ------  ------  ------
 Purchase price   10$     20$     30$     60$
 Sale price       15$     25$     35$     75$
 Profit           5$      5$      5$      15$

Thanks

Answer 1

You need a third table to aggregate purchase and sale prices for each item. Then, you compute profits and totals and rotate the rows so that each item becomes a column.

Check the script bellow (comments included) but remember that for large tables it might take a while to get the results.

--setup test variables
declare @table1 table(item varchar(50), purchase int)
declare @table2 table(item varchar(50), sale int)

insert into @table1 values ('item1', 10)
insert into @table1 values ('item2', 20)
insert into @table1 values ('item3', 30)

insert into @table2 values ('item1', 15)
insert into @table2 values ('item2', 25)
insert into @table2 values ('item3', 35)

--create a temporary table to hold each item's purchase price, sale price and profit
create table #sumTable (row varchar(50), item varchar(50), price int)

--insert purchase prices from table1
insert into #sumTable 
select 'purchase', item, purchase from @table1

--insert sale prices from table2
insert into #sumTable 
select 'sale', item, sale from @table2

--calculate profits for each item
insert into #sumTable
select 'profit', t1.item, sale - purchase
from @table1 t1
inner join @table2 t2 on t1.item = t2.item

--calculate totals
insert into #sumTable
select row, 'total', sum(price)
from #sumTable
group by row, item

--check what we have so far (uncomment if needed)
--select * from #sumTable

--roate the rows by row type, so that each item's name becomes a row
declare  @sql  nvarchar(max),
         @cols nvarchar(max)

select @cols = STUFF(
(select ',
SUM(CASE WHEN [item] = ''' + [item] + ''' THEN [price] ELSE NULL END) AS [' + [item] + ']'
from #sumTable
group by [item]
order by [item] 
for xml path(''),type).value('.','varchar(max)'), 1, 2, '')

set @sql = 'SELECT [row], ' + @cols + '  FROM #sumTable GROUP BY [row]'

execute(@sql)

--drop the temporary table
drop table #sumTable

The result looks like this

row         item1       item2       item3       total
----------- ----------- ----------- ----------- -------
profit      5           5           5           15
purchase    10          20          30          60
sale        15          25          35          75

Answer 2

You could first pivot your data, to mutate the data from rows to columns.

Then you union the results.

However, this looks like a reporting question - so the results may be better generated using a reporting tool such as SQL Server reporting services.