Incremental loads from SQL into any kind of Data Lake is easily achievable using many different methods. Where they fall apart is in replicating deletes, especially in large datasets, without having to do painful comparisons between source and target or imposing the significant overhead of full Change Data Capture.