How to join two tables in one view? - sql

Question

I am trying to create a view where I can see the items that have been planned to be shipped and have not been shipped, and the items have have been shipped but were not planned to.
In order to do this I have 2 tables with different data in them.
Table SC (actually shipped):
+---------+-----------------+----------------------+-------------+
| item_id | source_location | destination_location | shipped_qty |
+---------+-----------------+----------------------+-------------+
| 001 | California | South_Carolina | 80 |
+---------+-----------------+----------------------+-------------+
| 001 | California | South_Carolina | 0 |
+---------+-----------------+----------------------+-------------+
| 001 | California | Texas | 20 |
+---------+-----------------+----------------------+-------------+
| 003 | Texas | South_Carolina | 200 |
+---------+-----------------+----------------------+-------------+
| 004 | South_Carolina | Texas | 30 |
+---------+-----------------+----------------------+-------------+
| 004 | South_Carolina | Texas | 10 |
+---------+-----------------+----------------------+-------------+
Table SO (plan to ship items):
+---------+-----------------+----------------------+---------------+
| item_id | source_location | destination_location | planned_order |
+---------+-----------------+----------------------+---------------+
| 001 | California | South_Carolina | 100 |
+---------+-----------------+----------------------+---------------+
| 001 | California | South_Carolina | 100 |
+---------+-----------------+----------------------+---------------+
| 001 | California | Texas | 10 |
+---------+-----------------+----------------------+---------------+
| 003 | Texas | South_Carolina | 200 |
+---------+-----------------+----------------------+---------------+
| 004 | South_Carolina | Texas | 300 |
+---------+-----------------+----------------------+---------------+
| 004 | South_Carolina | Texas | 50 |
+---------+-----------------+----------------------+---------------+
So in this case, for example, since the item 001 has three different planned orders from California to South Carolina, I don't want it to show all the three orders in the view, I want it to be only in one row, but sum all the planned orders together, as showed below.
Desired Outcome:
+---------+----------------+-----------------+-------------+-------------+
| item_id | source_loc | destination_loc | shipped_qty | planned_qty |
+---------+----------------+-----------------+-------------+-------------+
| 001 | California | South_Carolina | 80 | 200 |
+---------+----------------+-----------------+-------------+-------------+
| 001 | California | Texas | 20 | 10 |
+---------+----------------+-----------------+-------------+-------------+
| 003 | Texas | South_Carolina | 200 | 200 |
+---------+----------------+-----------------+-------------+-------------+
| 004 | South_Carolina | Texas | 40 | 350 |
+---------+----------------+-----------------+-------------+-------------+
I have tried this so far:
SELECT o.source_location,
o.destination_location,
o.item_id,
o.planned_order,
c.shipped_qty
FROM SO_TRANSFER o, SC_TRANSFER c
But this hasn't worked since the shipped_qty does not match the item and this code also does not add the orders together.
By the way, I am using Microsoft SQL Server 2012.
Thank you!

Answer 1

I think you want:
select coalesce(s.item_id, p.item_id) as item_id,
coalesce(s.source_location, p.source_location) as source_location,
coalesce(s.destination_location, p.destination_location) as destination_location,
coalesce(s.shipped_qty, 0) as shipped_qty,
coalesce(planned_qty, 0) as planned_qty
from (select item_id, source_location, destination_location, sum(shipped_qty) as shipped_qty
from sc
group by item_id, source_location, destination_location
) s full join
(select item_id, source_location, destination_location, sum(planned_qty) as planned_qty
from so
group by item_id, source_location, destination_location
) p
on s.item_id = p.item_id and
s.source_location = p.source_location and
s.destination_location = p.destination_location;

Answer 2

You can try this:
SELECT A.item_id,
A.source_location AS source_loc,
A.destination_location AS destination_loc,
A.shipped_qty,
B.planned_order
FROM
(SELECT item_id,source_location,destination_location,SUM(shipped_qty)
AS shipped_qty
FROM SC GROUP BY item_id,source_location,destination_location) A,
(SELECT item_id,source_location,destination_location,SUM(planned_order)
AS planned_order
FROM SO GROUP BY item_id,source_location,destination_location) B
WHERE A.item_id = B.item_id AND
A.source_location= B.source_location AND
A.destination_location= B.destination_location
EDIT: I just realize my answer is similar to Gordon Linoff's answer, and his answer got more feature such as handling data that exists in one table only using COALESCE in T-SQL and FULL JOIN property. Since I worked 1 hour for this answer, so I will just leave it here.

Answer 3

The point is you should SUM(quantity) from each table first, then you can easy JOIN 2 table with condition:
ON so.item_id = sc.item_id AND so.source_loc = sc.source_loc AND so.destination_loc = sc.destination_loc

Answer 4

The following query will do what you are looking for
Create table #SC(
item_id varchar(50),
source_location varchar(max),
Destination_location varchar(max),
Shipped_qty int)
insert into #SC values('001','california','sourth_carolian',80)
insert into #SC values('001','california','sourth_carolian',0)
insert into #SC values('001','california','Texas',20)
insert into #SC values('003','Texas','sourth_carolian',200)
insert into #SC values('004','sourth_carolian','Texas',30)
insert into #SC values('004','sourth_carolian','Texas',10)
--select * from #SC
Create table #SO(
item_id varchar(50),
source_location varchar(max),
Destination_location varchar(max),
Planned_order int)
insert into #SO values('001','california','sourth_carolian',100)
insert into #SO values('001','california','sourth_carolian',100)
insert into #SO values('001','california','Texas',10)
insert into #SO values('003','Texas','sourth_carolian',200)
insert into #SO values('004','sourth_carolian','Texas',300)
insert into #SO values('004','sourth_carolian','Texas',50)
--select * from #SO
select C.item_id,C.source_location,C.Destination_location, sum(C.Shipped_qty) as Shipped_qty, po.planned_order from #SC C
outer apply
(select sum(Planned_order) as planned_order from #SO
where source_location+Destination_location=C.source_location+C.Destination_location
group by item_id,source_location,Destination_location ) as PO
group by C.item_id,C.source_location,C.Destination_location,po.planned_order

Answer 5

You can create a SELECT statement, and use that as a table in the FROM clause:
SELECT o.source_location,
o.destination_location,
o.item_id,
o.planned_order,
c.shipped_qty_sum
FROM SO_TRANSFER o
INNER JOIN (SELECT SUM(shipped_qty) AS shipped_qty_sum,
source_location,
item_id
FROM SC_TRANSFER
GROUP BY source_location, item_id) c
ON o.item_id = c.item_id AND o.source_location = c.source_location