I'm trying to create a retail data set with the following basic constructs
* Sales transactions
This works semi-well. I do get unique rows etc for everything, using a join to both item and customer. However, a customer usually do not buy only one (1) item, but the shopping basket generally hold 1-50 items of different quantities.
So, what I would like to achieve, is to set the item_id-join, to bring back a random of 1-50 items and then repeat the customer_id and the purchase_date, ie imitating a real shopping scenario.
Is this possible? (hardcoded scenario distribs are out of the question, I have a 10000 item data set that should be randomized)