What is the best practice while loading delta table , do I infer the schema or provide the schema?

Data Engineering

Join discussions on data engineering best practices, architectures, and optimization strategies within the Databricks Community. Exchange insights and solutions with fellow data engineers.

I am loading avro files into the detla tables. I am doing this for multiple tables and some files are big like (2-3GB) and most of them are small like in few MBs.

I am using autoloader to load the data into the delta tables.

My question is:

What is the best practice , do I infer the schema or provide the schema while using autoloader.

I understand that if I provide the schema , the processing would be faster however it would be time consuming to provide the schema for tables that has many columns.

Any suggestions?