what is best practice to handle the concurrency issue in batch processing?

andrew0117
Contributor

Normally, our ELT framework takes in batches one by one and loads the data into target tables. But if more than one batches come in at the same time, the framework will break due to the concurrency issue that multiple sources are trying to write the same tables at the same time.

what is the best way or standard procedure to solve this concurrency issue? Thanks!

jose_gonzalez
Databricks Employee
Databricks Employee

you can partition you table to avoid the changes of getting this exception.