Sql query, simple table, big data why long duration?

Question

SQL

There is a table of actual weather service data with 2 columns: primary key datetime2 and varbinary
The usual "select from where date between" gives ~ 1 million rows per month and lasts ~ 10 seconds.
Given that the sample can be for a year, for the task at hand it takes a very long time.
How can the sampling time be reduced?
PS: nolock is worth it, rebuilding the indexes did not give any result

Answer 1

I would go the other way. Getting 1 million rows - this is clearly not the end result, what do you want to do with them? Can write such request which will give out at once the final result?