Optimize Performance DP-700 Practice Questions (Page 5)

Questions 21–25

1. 21application · medium

   A pipeline uses a Spark job to process data. The job fails intermittently due to executor memory issues. The engineer wants to improve reliability without changing the code. What should you do?

   • AIncrease the executor memory in the Spark pool configuration.
   • BReduce the number of partitions in the job.
   • CDisable checkpointing to reduce overhead.
   • DUse a smaller file format like CSV.
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A Lakehouse table is frequently queried with filters on a date column. The table is partitioned by a different column. Queries are slow because they scan all partitions. What should you do?

   • AUse V-order on the table.
   • BIncrease the number of partitions.
   • CChange the partition column to the date column.
   • DCreate a materialized view on the table.
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   A lakehouse table in Delta format is written by a streaming job every few minutes, resulting in many small files. Queries that scan the table are slow because of the high file count. The table is also used by a nightly batch job that reads the entire table. Which optimization should the engineer schedule to run periodically?

   • AIncrease the number of streaming partitions to write more files.
   • BRun VACUUM to delete old files and reduce storage.
   • CRun OPTIMIZE on the table to compact small files into larger ones.
   • DDisable the streaming job and use a batch job instead.
   Reveal Solution Discussion Select an answer first
5. 24application · medium

   A Spark job reads a large Delta table, performs a join with a small lookup table, and writes the result to a new Delta table. The job is slow because of a large shuffle. The lookup table is small enough to fit in memory on each executor. Which optimization should the engineer apply?

   • ACache the large table in memory to avoid reading it again.
   • BUse a broadcast hint on the lookup table to avoid shuffling the large table.
   • CUse a shuffle join to ensure correct results.
   • DIncrease the number of shuffle partitions to reduce partition size.
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A Fabric warehouse contains a large fact table with billions of rows. Users report that a common reporting query that aggregates sales by region and month is slow. The query filters on a region column and a sales_date column. The warehouse uses default distribution and indexing. Which optimization should the engineer apply to improve query performance?

   • APartition the table by region and create a view that pre-aggregates the data.
   • BUse a heap table with round-robin distribution to avoid index maintenance overhead.
   • CCreate a nonclustered index on the sales_date column only.
   • DCreate a clustered columnstore index on the fact table and use hash distribution on the region column.
   Reveal Solution Discussion Select an answer first