Ipolate/Epolate?

[Johan Verburg]

Hello Statalist,

I am using Stata 14. I have a dataset with over 1000 products and want calculate some price indices.
I want to use the starting quantity of each product and use that to make some calculations. When the product is available at the starting period I can use ipolate to set the quantity for that product as quantity0. When the starting period has no quantity available this method does not work anymore. Below is a subset shown with some data

clear all
input id product_id period price quantity
1 1 1 3 4
2 1 2 2 5
3 1 3 3 3
4 1 4 2 6
5 2 1 . . 
6 2 2 . .
7 2 3 4 3
8 2 4 5 2
9 3 1 . .
10 3 2 1 2
11 3 3 2 5
12 3 4 2 3
end

local start period[1]
gen temp = quantity
replace temp = . if period != `start'
ipolate temp product_id, generate(quantity0)

Question: How can I get a variable quantity0 that has a value of 4 for all periods for product_id == 1,
a value of 3 for periods 3 and 4 for product_id == 2,
and a value of 2 for periods 2, 3 and 4 for product_id == 3.

Thank you in advance

[Nick Cox]

If any interpolation makes sense here, it would be

ipolate quantity period, by(product_id) generate(something)

What rule is that which yields

4 for all periods for product_id == 1,
3 for periods 3 and 4 for product_id == 2,
2 for periods 2, 3 and 4 for product_id == 3.

?

[Johan Verburg]

4 is the starting quantity for product_id ==1
3 is the starting quantity for product_id ==2
2 is the starting quantity for product_id ==3

So I am looking for a way to get the following result:

quantity0
4
4
4
4
.
.
3
3
.
2
2
2