Group By first day of month and join with a separate table - sql

Question

I have 2 tables in SQL
one with monthly sales targets:
Date Target
1/7/17 50000
1/8/17 65000
1/9/17 50000
1/10/17 48000
etc...
the other with sales orders:
TxnDate JobNum Value
3/7/17 100001 20000
3/7/17 100002 11000
8/7/17 100003 10000
10/8/17 100004 15000
15/9/17 100005 20000
etc...
what I want is a table with following:
Date Target Sales
1/7/17 50000 41000
1/8/17 65000 15000
1/9/17 50000 20000
please help me I'm a newbie to coding and this is doing my head in.. :)

Answer 1

Assuming your 1st table is targetSales and your 2nd table is Sales and your database is SQL Server:
select
t.date
, t.target
, isnull(sum(s.value), 0) as Sales
from targetSales t
left join Sales s
on (month(t.date) = month(s.date)
and year(t.date) = year(s.date))
group by t.date
, t.target
You can follow a similar approach if you use a different database, just find the equivalents of month() and year() functions for your RDBMS.

Answer 2

try this
select tb1.date,tb1.target,tb2.value from table1 as tb1
INNER JOIN (select sum(value) as sales, date from table2 group by date) as tb2
on tb1.date = tb2.date,
you can use this script for daily targets

Answer 3

An another way around, looks like in target table the date is always the first day of the month. So in the sales table, just round the TxnDate column value to first day of the month.
Query
select t1.[date],
max(t1.[target]) as [target],
coalesce(sum(t2.[value]), 0) as [value]
from [targets] t1
left join [sales] t2
on t1.[Date] = dateadd(day, - datepart(day, t2.[txnDate]) + 1, t2.[txnDate])
group by t1.[Date];
demo

Answer 4

If you take any datetime value in SQL Server, calculate the number of months from that date to zero datediff(month,0,TxnDate) then add that number of moths to zero dateadd(month, ... , 0) you get the first day of the month for the original datetime value. This works in all versions of SQL Server. With this we can sum the values of the orders by the first day of the month, then join to targets using that date.
CREATE TABLE Orders
([TxnDate] datetime, [JobNum] int, [Value] int)
;
INSERT INTO Orders
([TxnDate], [JobNum], [Value])
VALUES
('2017-07-03 00:00:00', 100001, 20000),
('2017-07-03 00:00:00', 100002, 11000),
('2017-07-08 00:00:00', 100003, 10000),
('2017-08-10 00:00:00', 100004, 15000),
('2017-09-15 00:00:00', 100005, 20000)
;
CREATE TABLE Targets
([Date] datetime, [Target] int)
;
INSERT INTO Targets
([Date], [Target])
VALUES
('2017-07-01 00:00:00', 50000),
('2017-08-01 00:00:00', 65000),
('2017-09-01 00:00:00', 50000),
('2017-10-10 00:00:00', 48000)
;
GO
9 rows affected
select dateadd(month,datediff(month,0,TxnDate), 0) month_start, sum(Value) SumValue
from Orders
group by dateadd(month, datediff(month,0,TxnDate), 0)
GO
month_start | SumValue
:------------------ | -------:
01/07/2017 00:00:00 | 41000
01/08/2017 00:00:00 | 15000
01/09/2017 00:00:00 | 20000
select
t.[Date], t.Target, coalesce(o.SumValue,0)
from targets t
left join (
select dateadd(month,datediff(month,0,TxnDate), 0) month_start, sum(Value) SumValue
from Orders
group by dateadd(month, datediff(month,0,TxnDate), 0)
) o on t.[Date] = o.month_start
GO
Date | Target | (No column name)
:------------------ | -----: | ---------------:
01/07/2017 00:00:00 | 50000 | 41000
01/08/2017 00:00:00 | 65000 | 15000
01/09/2017 00:00:00 | 50000 | 20000
10/10/2017 00:00:00 | 48000 | 0
dbfiddle here

Answer 5

This is not the best solution but this will give you a correct result.
select date,target,(
select sum(value)
from sales_orders s
where datepart(m,s.TxnDate) = datepart(m,targets.Date)
and datepart(year,s.TxnDate) = datepart(year,targets.Date)
) as sales
from targets