Question

MySQL is still pretty new to me, so I’m having difficutly finding the answer to this particular problem.

I have roughly 60 vendor tables that I would like to join into one table. Alot of our vendors carry the same items so I would like to add all of the items to one table, and those that have the same sku, add the quantities together. I would also like to insert the greater cost and price, and round it up to .99, but if it gets too complicated or can’t be sone, I will stick with just the sku’s and quanitity for now. I would like to start with 2 tables until I can figure it out, then once I grasp the solution I can work on all of them. Here’s an example of 2 tables:

Table A (SKU will be the primary key)
SKU     Vendor A SKU         Cost     Price      QTY
1000        VA100            49.99    62.99      10
1001        VA101            64.44    81.99      6
1002        VA102            37.67    47.99      24
1003        VA103            28.33    35.99      16
1004        VA104            26.33    35.99      18

Table B (SKU will be the primary key)

SKU     Vendor B SKU         Cost    Price       QTY
1000        VB100            49.99   62.95        21
1001        VB101            64.44   81.95        16
5001        VB501            32.67   40.95         4
5002        VB502            88.44   110.95      12
5003        VB503            87.21   109.95       2

So the End Result will look like this

SKU     Vendor A SKU     Vendor B SKU      Cost           Price         QTY
1000        VA100               VB100      49.99         62.99          31 <-- Sum
1001        VA101               VB101      64.44         81.99          22 <-- Sum
1002        VA102                          37.67         47.99          24
1003        VA103                          28.33         35.99          16
1004        VA104                          26.33         35.99          18
5001                            VB501      32.67         40.95           4
5002                            VB502      88.44         110.95         12
5003                            VB503      87.21         109.95          2

As you can see the first 2 had the quantities added together, as well as the greater of cost and price inserted. I’ve tried variations of:

Create Table MASTER
SELECT TableA.sku,TableA.VendorA_sku,TableA.cost,TableA.price,TableA.qty
UNION ALL
SELECT TableB.sku,TableB.VendorB_sku,TableB.cost,TableB.price,TableB.qty
GROUP BY sku

And variations of:

Create Table MASTER
AS SELECT TableA.sku,TableA.VendorA_sku,TableA.cost,TableA.price,SUM(TableA.qty+TableB.qty)
GROUP BY sku

I’ve tried numerous different types of Joins and unions and can’t seem to get anything to look the way I need it to, nor have I played around with the adding and rounding of cost and price, as I can’t get past the adding of quanitity in duplicate Primary keys. I have to do this on a daily basis with roughly 60 tables and about 600k sku’s, so any help would be greatly appreciated.

(I also Apologize in advance if I didn’t type this correctly as I’m new to these types of forums)

Answer 1

The following should work for you:

select sku,
  min(case when Type = 'VendorA' THEN Vendor END) VendorA_SKU,
  min(case when Type = 'VendorB' THEN Vendor END) VendorB_SKU,
  cost,  
  price, 
  sum(QTY) Total
from
(
  select sku, VendorA_SKU Vendor, cost, price, qty, 'VendorA' Type
  from tablea
  union all 
  select sku, VendorB_SKU, cost, price, qty, 'VendorB' Type
  from tableb
) x
group by sku

see SQL Fiddle with Demo

A few things to decide, do you want the AVG() or MAX() cost and price if the values are different or which one do you want to return? If you do then you will use the following to return MAX():

select sku,
  min(case when Type = 'VendorA' THEN Vendor END) VendorA_SKU,
  min(case when Type = 'VendorB' THEN Vendor END) VendorB_SKU,
  max(cost) MaxCost,
  max(price) MaxPrice,
  sum(QTY) Total
from
(
  select sku, VendorA_SKU Vendor, cost, price, qty, 'VendorA' Type
  from tablea
  union all 
  select sku, VendorB_SKU, cost, price, qty, 'VendorB' Type
  from tableb
) x
group by sku

see SQL Fiddle with Demo