On what basis was partitioning done in Amazon Athena?

Partitioning in Amazon Athena is based on columns that are frequently used in query filters to reduce the amount of data scanned and improve performance.

Most often, partitioning is done by dates (for example, year, month, day) since many analytical queries filter data by time. Other logical columns such as region, category, or status can also be used if they allow for effective restriction of the dataset.

Example:

CREATE EXTERNAL TABLE logs (
  id string,
  event string,
  ...
)
PARTITIONED BY (year string, month string, day string)
LOCATION 's3://bucket/path/';

Partitioning is based on analyzing query patterns and data structure to minimize the scanning of unnecessary files.