SQL was designed to work on structured data—think tables with well-defined
columns and rows, much like an Excel spreadsheet. Much of the power of SQL
comes from the power of the underlying database engine and the optimizer. Many
people use databases running on powerful computers, without ever thinking about
the underlying hardware. That is the power of SQL: The same query that runs on a
mobile device can run on the largest grid computer, taking advantage of all
available hardware with no changes to the query.
The part of the SQL language used for analysis is the SELECT statement. Much of
the rest of the language is about getting data in to databases. Our concern is
getting information out of them to solve business problems. The SELECT statement
describes what the results look like, freeing the analyst to think about what to do,
instead of how to do it.
SQL (when used for querying) is a descriptive language rather than a
procedural language. It describes what needs to be done, letting the SQL
engine optimize the code for the particular data, hardware, and database
layout where the query is running, and freeing the analyst to think more about
the business problem.