How to use rank() function ?

Scenario. Rank the companies by country based on their turnover in Hive ?
companyName, country, turnover

Solution
Hive>
SELECT companyName, country, turnover,
       rank() over (PARTITION BY country ORDER BY turnover DESC) as rank
FROM sales;

SORT BY vs. ORDER BY

• ORDER BY performs a total ordering of the query result set
• All the data is passed through a single reducer
• takes long time to execute for larger data

• SORT BY sorts the data within each Reducer - Local ordering
• Each each reducer’s output will be sorted
• doesn't achieve a total ordering on the dataset