I have a panel data set that is of the form
shpro date price
1 1 100
1 2 100
1 3 100
1 4 100
2 1 98
2 2 100
2 3 102
2 4 104
3 1 99
3 2 100
etc. etc. etc.
where shpro is a variable that represents the same product within the same shop ie it is product-shop specific. The date represents the date of the price reading and price is price of the product in that shop.

I am carrying out a fixed effects regression with time fixed effects and shop-product fixed effects. I wish to condition my regression on the fact that the price of each product at date=1 is different to the price of that same product at date=4. I initially generated a variable for price at date=1 and date=4 but of course this does not work since each observation has only one date. I have an inkling that I may need to reshape the data, but I am not entirely sure on how to do this.

Any help will be so much appreciated.