DATABRICKS-DEP Question 60

Community Discussion

Written by other learners. Comments can disagree with the marked answer — check the official documentation when in doubt.

• Chose D0

  - A watermark tells Spark how long it must keep state to recognize late duplicates, and streaming deduplication uses that state to suppress repeats. Micro-batch-only `DISTINCT` logic does not protect against duplicates that arrive in later batches.