Utilize Common Data Sources Such as JDBC, Files, Etc., to Efficiently Read from and Write to Spark DataFrames Using Spark SQL, Including Overwriting and Partitioning by Column. ASSOCIATE-DEVELOPER-APACHE-SPARK Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   Which Spark SQL format string is used to read a directory of Parquet files into a DataFrame?

   • Ajson
   • Bcsv
   • Cparquet
   • Dorc
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   Which Spark SQL option is commonly used when reading a CSV file to treat the first line as column names?

   • AinferSchema
   • Bdelimiter
   • Cquote
   • Dheader
   Reveal Solution Discussion Select an answer first
4. 18expert · hard

   A team needs to read a large table from a JDBC source. The table has a primary key column 'id' that is not uniformly distributed (e.g., many rows have id values clustered in a small range). They want to parallelize the read but are concerned about data skew causing some tasks to take much longer than others. Which approach is the most robust?

   • AUse partitionColumn="id", lowerBound=1, upperBound=1000000, numPartitions=100.
   • BUse a custom query with a WHERE clause that splits the data into balanced ranges based on a different, uniformly distributed column.
   • CRead the table without partitioning and rely on Spark's default parallelism.
   • DUse partitionColumn="id", lowerBound=1, upperBound=1000000, numPartitions=10.
   Reveal Solution Discussion Select an answer first
5. 19foundation · easy

   What is the default behavior of Spark's overwrite mode when writing to a file-based data source?

   • AIt fails if the target path already contains data.
   • BIt appends the new data to the existing data.
   • CIt writes the new data to a temporary directory and then renames it.
   • DIt deletes the existing data at the target path before writing the new data.
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A company runs a daily ETL job that writes a DataFrame of customer updates to a JDBC-backed table. The table already exists and contains yesterday's data. The job must replace the entire contents of the table with today's data, and the table must not contain any rows from the previous run. Which write mode should they use?

   • Aignore
   • Boverwrite
   • Cerrorifexists
   • Dappend
   Reveal Solution Discussion Select an answer first