Fractal Analytics Engineer Interview Questions, Process, and Tips for Experienced

First round was coding round that comprised of 4 questions, 1 sql and 3 programming questions. Out of 3, if you are able to run 2 code successfully, you'll qualify for the next round

Q1. In this round, interviewer will ask basic interview questions first, then they'll give multiple sql questions, few questions based on spark and few programming questions
Add your answer

Q1. Difference between Spark and MapReduce. Spark joins like broadcast and sort merge
Ans. 
Spark is faster than MapReduce due to in-memory processing and DAG execution.
• Spark uses DAG (Directed Acyclic Graph) execution while MapReduce uses batch processing.

• Spark performs in-memory processing while MapReduce writes to disk after each operation.

• Spark has a more flexible programming model with support for multiple languages.

• Spark has built-in libraries for machine learning, graph processing, and stream processin...read more
Answered by AI
Add your answer
Q2. Spark file formats like parquet, orc, avro etc. Which one you are using and why
Add your answer
Q3. Spark optimization techniques
Ans. 
Spark optimization techniques improve performance and efficiency of Spark applications.
• Partitioning data to reduce shuffling

• Caching frequently used data

• Using broadcast variables for small data

• Tuning memory allocation and garbage collection

• Using efficient data formats like Parquet

• Avoiding unnecessary data shuffling

• Using appropriate hardware configurations

• Optimizing SQL queries with appropriate indexing and partitioning
Answered by AI
Add your answer
Q4. Basic programming language based questions
Add your answer
Q5. Hive- partition vs bucketing
Ans. 
Hive partitioning is dividing data into smaller, manageable parts while bucketing is dividing data into equal parts based on a hash function.
• Partitioning is useful for filtering data based on a specific column

• Bucketing is useful for evenly distributing data for faster querying

• Partitioning can be done on multiple columns while bucketing is done on a single column

• Partitioning creates separate directories for each partiti...read more
Answered by AI
Add your answer
Q6. Hive optimization techniques
Ans. 
Hive optimization techniques improve query performance and reduce execution time.
• Partitioning tables to reduce data scanned

• Using bucketing to group data for faster querying

• Using vectorization to process data in batches

• Using indexing to speed up lookups

• Using compression to reduce storage and I/O costs
Answered by AI
Add your answer
Q7. Recent projects you've worked on
Add your answer
Q8. SQL based questions, window functions
Add your answer

Q1. Why do you want to switch
Add your answer
Q2. Have you ever faced any problem in previous projects, how you handled that
Add your answer