Optimize and Troubleshoot Data Storage and Data Processing DP-203 Practice Questions (Page 8)

Questions 36–38

1. 36application · medium

   A Spark job that joins a large fact table with a small dimension table is running slowly. The fact table is skewed on the join key 'RegionID', with one region containing 70% of the rows. The dimension table has 50 rows. Which technique should the data engineer use to improve performance?

   • AIncrease the number of shuffle partitions to 500.
   • BUse a sort-merge join instead of a broadcast join.
   • CAdd a filter to exclude the hot region.
   • DBroadcast the dimension table and salt the fact table's join key.
   Reveal Solution Discussion Select an answer first
2. 37foundation · easy

   A Spark job fails with an 'OutOfMemoryError' in the executor logs. Which initial troubleshooting step is most appropriate?

   • ACheck the executor memory configuration and the amount of data processed per task.
   • BImmediately increase the number of executors without changing memory settings.
   • CDisable dynamic allocation to keep executors constant.
   • DSwitch the cluster to a different runtime version.
   Reveal Solution Discussion Select an answer first
3. 38foundation · easy

   Which approach helps prevent data spill in a Spark job?

   • AUse a larger file format like CSV instead of Parquet.
   • BDecrease the number of partitions to reduce task overhead.
   • CDisable compression to reduce CPU overhead.
   • DIncrease the number of partitions to reduce the amount of data each task processes.
   Reveal Solution Discussion Select an answer first