SparkJun
Databricks Employee
Databricks Employee

 

If you use the applyChanges method in DLT for Change Data Capture (CDC), you can delete records manually without affecting the consistency of the table, as applyChanges respects manual deletions. You must configure your DLT pipeline to respect manual deletes. This involves ensuring that your pipeline logic does not reintroduce the deleted records during the next run. For instance, if your pipeline is set to always overwrite the entire table (e.g., through full refreshes), manual deletions will not be retained unless the table is configured to allow manual updates/deletes.