Which PySpark expression returns the total number of records in the fact table grouped by ProductKey, with results sorted by count in descending order?

Which PySpark expression returns the total number of records in the fact table grouped by ProductKey, with results sorted by count in descending order?

• A. df.groupBy("ProductKey").count().sort("count", descending=True).show()
• B. df.groupBy("ProductKey").count().sort("count", ascending=True).show()
• C. df.groupBy("Key").count().sort("count", descending=True).show()
• D. df.groupBy("ProductKey").count().orderBy("ProductKey").show()

Answer: (A)

Grouping by ProductKey and counting gives the total number of records for each product key. After performing the groupBy and count, you have a column named count that holds the number of records per ProductKey. Sorting by that count in descending order lists the products from the highest to the lowest total. The expression df.groupBy("ProductKey").count().sort("count", descending=True).show() does exactly this: it groups by ProductKey, counts the records in each group, and sorts the results by the count in descending order so the most frequent product keys appear first. Sorting by count ascending would be the opposite and not meet the requirement. Grouping by a different column would produce counts per the wrong key. Sorting by ProductKey would order the results by the key values themselves, not by how many records each key has.

Grouping by ProductKey and counting gives the total number of records for each product key. After performing the groupBy and count, you have a column named count that holds the number of records per ProductKey. Sorting by that count in descending order lists the products from the highest to the lowest total.

The expression df.groupBy("ProductKey").count().sort("count", descending=True).show() does exactly this: it groups by ProductKey, counts the records in each group, and sorts the results by the count in descending order so the most frequent product keys appear first.

Sorting by count ascending would be the opposite and not meet the requirement. Grouping by a different column would produce counts per the wrong key. Sorting by ProductKey would order the results by the key values themselves, not by how many records each key has.