What do you do in SQL vs Pandas?

My work primarily stores data in a full databases. Pandas has a lot of similar functionality to SQL in regards to the ability to group data and preform calculations, even being able to take full on SQL queries to import data. Do you guys do all your calculations in the query itself, or in python after the data has been imported? What about with grouping data?

data databases datascience import pandas query sql sql queries work