What is the best file format for a temporary table?

brickster_2018
Databricks Employee
Databricks Employee

‎06-23-2021 02:14 PM

As part of my ETL process, I create intermediate/staging temporary tables. These tables created are read at a later point in the ETL and finally cleaned up. Should I use Delta? Using Delta creates the overhead of running optimize jobs, which would delay my ETL. Enabling optimize writes also will add delay to my ETL

What is the best practice/recommendation here.

brickster_2018
Databricks Employee
Databricks Employee

‎06-24-2021 06:38 AM

  • The temporary table can be created using Delta, although enabling optimize writes on the temporary table can increase the read performance in the subsequent stage. Overall it's a good practice to use Delta.
  • Eventual consistency-related issues are common when dealing with temporary tables. Using Delta can help to address those issues.

View solution in original post

Sebastian
Contributor

‎09-13-2021 05:54 AM

Agree.. the intermediate delta tables helps since it brings reliability to the pipeline.

0 Kudos