I have a table with this schema:
Fruit Truck ID Bucket ID Date
------ ----- --------- ----------
Apple 1 101 2018/04/01
Apple 1 101 2018/04/10
Apple 1 112 2018/04/16
Apple 2 782 2018/08/18
Apple 2 782 2018/09/12
Apple 1 113 2019/09/12
Apple 1 113 2019/09/21
My goal is to write an SQL script that returns the start and end dates of each truck & bucket pair for each fruit. The intended result is below:
Fruit Truck ID Bucket ID Start Date End Date
------ ----- --------- ---------- ----------
Apple 1 101 2018/04/01 2018/04/16
Apple 1 112 2018/04/16 2018/08/18
Apple 2 782 2018/08/18 2018/09/12
Apple 1 113 2019/09/12 2019/09/21
I have tried solving this through lag/lead window functions, but it the dates are not correct. Is there another method of solving this using window functions or do I have to create sub queries for this?
I think you want aggregation and window functions:
select fruit, truck_id, bucket_id,
min(date) start_date,
lead(min(date), 1, max(date)) over(partition by fuit order by min(date)) end_date
from mytable
group by fruit, truck_id, bucket_id