DATABRICKS-DEP Question 87

Community Discussion

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

• Chose C1

  For wide transformations (shuffles), you want: More executors = Better shuffle parallelism More network interfaces = Higher aggregate bandwidth Distributed I/O = Parallel shuffle reads/writes. 16x network bandwidth: Shuffle data distributed across 16 network interfaces ✅ True parallelism: 16 nodes simultaneously reading/writing shuffle data ✅ Better fault tolerance: Losing 1 executor = only 6% of work lost vs 100% ✅ Reasonable cores/executor: 10 cores (not ideal, but better than 160!)

• Chose B1

  suretly B

• Chose B1

  merge will work if no match