How to emulate temporary tables in Athena

Use CTAS to create a temporary table in Athena

Bartosz Mikulski 28 Dec 2020 – 1 min read

Currently (I wrote this article in October 2020), Athena does not support temporary tables, but we can easily emulate them using the CREATE TABLE AS SELECT statements:

CREATE TABLE some_temp_table
    WITH (format = 'PARQUET')
    AS SELECT column_A, column_B, column_C
    FROM source_table;

Unfortunately, we have to remember about removing the table when we no longer need it.

DROP TABLE some_temp_table

If we don’t specify the S3 location, Athena will use the default results bucket as the storage location. I think it is good enough in the case of a temporary table.

Note that I used Parquet as the storage file type. In general, you should pick a file format that is best for the operations you want to perform later. Parquet is a columnar storage file that stores metadata about the content to scan and find the relevant data quickly.

Because the table we create is just a regular table, we can also use partitioning:

CREATE TABLE some_temp_table
    WITH (
        format = 'PARQUET',
        partitioned_by = ARRAY['column_A'])
    )
    AS SELECT column_A, column_B, column_C
    FROM source_table;

How to emulate temporary tables in Athena

How to enable S3 bucket versioning using Terraform

Send SMS from an Airflow DAG using AWS SNS

How to emulate temporary tables in Athena

How to enable S3 bucket versioning using Terraform

Send SMS from an Airflow DAG using AWS SNS

Related Posts

How to use AWSAthenaOperator in Airflow to verify that a DAG finished successfully