Urgent Help Needed - Databricks Notebook Failure Handle for Incremental Processing

rrajan
New Contributor II

I have created a notebook which helps in creating three different gold layer objects from one single silver table. All these tables are processed incremently. I want to develop the failure handling scenario in case if the pipeline fails after loading few of the records in the first table or if one of the gold table loaded successfully and the second failed.

In this case while re-running the pipeline from scratch I don't want to insert already inserted records again in one of  the gold table. How to handle this type of scenario ?