CREATE TABLE USING

Using this SQL technique, you define the table schema, and create the table directly from the CSV file without switching to PySpark for initial data loading and df creation etc.

-- Drop the table if it exists
DROP TABLE IF EXISTS hollywood;

-- Create an external table from the CSV file
CREATE TABLE hollywood (
  movieName STRING,
  actor STRING
)
USING CSV
OPTIONS (
  path '/mnt/movies.csv',
  header 'true',
  inferSchema 'true'
);