The data for this test is pretty simple and fairly standard sales
data. There are two tables, although in reality there could be many more. The
Sales Table could have millions of rows. A fact were are not considering in these
tests, but one you need to consider. So there are two tables - a sales table and a
model table. The sales table has information on sales. The model table maps
SKUs to product lines. The tables are laid at below.
| Sales Table |
| Date |
| SKU |
| Sales Units |
| Sales Dollars |
| Sales Margin |
|
| SKU Table |
| SKU |
| Alpha SKU Description |
| Product Line |
|
Let's go to the first test.
Period to Period comparisons.
First some quick questions.
- Is all your data in relational form?
- Is it 'clean' and easy to work with, no obscure codes that if it's a z
then something else if it's an h then something else.
- Do you have plenty of DBA resources available?
- What about Extract, Transform and Load. Is that a bottleneck?