Spark interview questions for Senior Developers

Ashutosh Kumar
3 min readMay 8, 2020

--

Below is the compiled and tailored list of questions one can encounter in their Spark Interviews :

Before this, the first question one should be well aware of “Why spark is needed ? “.

1. Why Spark is Faster Than Hadoop? Hadoop Vs spark
2. Which language to choose and Why? Scala vs Python
3. Explain about the Apache Spark Architecture
4. What do you understand by Spark Execution Model
5. Brief about spark internals, Spark Session vs Spark Context
6. Spark Driver vs Spark Executor
7. Executor vs Executor core
8. Yarn client mode vs cluster mode
9. What is RDD and what do you understand by partitions?
10. What do you understand by Fault tolerance in Spark?
11. Spark vs Yarn Fault tolerance
12. Why Lazy evaluation is important in Spark?
13. Transformations vs actions
14. Map vs FlatMap
15. Spark Map vs Map Partition
16. Wide vs Narrow transformations
17. Reduce by key vs Group by key
18. What do you understand by Spark Lineage
19. Spark Lineage vs Spark DAG
20. Spark cache vs Spark persist
21. What do you understand by AggregateByKey and CombineByKey?
22. Briefly explain about Spark Accumulator
23. What do you mean by Broadcast variables?
24. Spark UDF functions, Why one should avoid UDF?
25. Why one should avoid RDDs, what is the alternative?
26. What are the benefits of a data frame?
27. What do you understand by Vectorized UDF?
28. Which one is better and when you should use, RDDs, Dataframe and Datasets?
29. Why Spark Dataset is typesafe?
30. Explain about Repartition and Coalesce.
31. How to read JSON from Spark?
32. Explain about Spark WIndow functions and it’s usage.
33. Spark Rank vs Dense Rank
34. Partitions vs Bucketing
35. Explain about catalyst optimizer
36. Stateless vs Stateful transformations
37. StructType and StructField
38. Explain about Apache parquet
39. What do you understand by CBO, Spark Cost Based Optimizer?
40. Explain Broadcast variable and shared variable with examples
41. Have you ever worked on Spark performance tuning and executor tuning
42. Explain Spark Join without shuffle
43. Explain about Paired RDD
44. Cache vs Persist in Spark UI
45. Why one should avoid groupBy?
46. How to decide the number of partitions in a data frame?
47. What is DAG? Explain in details.
48. Persistence vs Broadcast in Spark
49. Partition pruning and predicate pushdown
50. Fold vs reduce in Spark
51. Explain the interlinking of Pyspark and Apache Arrow
52. Explain about bucketing in Spark SQL
53. Explain dynamic resource allocation in Spark
54. Why fold-left and fold-right are not supported in Spark?
55. How to decide the number of executors and memory for any spark job?
56. Different types of cluster managers in spark
57. Can you explain how to minimize data transfers while working with Spark?
58. What are the different levels of persistence in Spark?
59. What is the function of filer()?
60. Define Partitions in Apache Spark?
61. What is the difference between reducing () and take() function?
62. Define YARN in Spark?
63. Can we trigger automated clean-ups in Spark?
64. What is another method than “Spark.cleaner.ttl” to trigger automated clean-ups in Spark?
65. What is the role of Akka in Spark?
66. Define SchemaRDD in Apache Spark RDD
67. What is a Spark Driver?

--

--

Ashutosh Kumar
Ashutosh Kumar

Written by Ashutosh Kumar

Backend Engineering | BIT Mesra | Building Microservices and Scalable Apps | Mentor