Databricks Certified Data Engineer - Associate
Get started today
Ultimate access to all questions.
A data engineer is curating data in the silver layer of a hospital management data warehouse system. The data engineer is trying to aggregate hospital billing data from a table patient_billing to generate a daily revenue fact table daily_revenue.
Assume this as a sample of dataframe billing_df:
| billing_id | patient_id | department | billing_date | amount_billed | quantity |
|---|---|---|---|---|---|
| 401 | p001 | Cardiology | 2024-03-01 | 1500 | 1 |
| 402 | p002 | Radiology | 2024-03-02 | 3000 | 1 |
| 403 | p001 | Cardiology | 2024-03-01 | 6500 | 1 |
| 404 | p003 | Radiology | 2024-03-03 | 500 | 1 |
Which code snippet aggregates the amount billed per day with the unique invoices from a Dataframe billing_df? (Options were provided separately.)
Other Sources
Community
DDucse
Explanation:
Correct answer: D.
Reasoning:
- Goal: calculate total daily revenue (sum of amount_billed per billing_date) and the number of unique invoices per day (unique billing_id count).
- Option A wrongly uses sum("billing_id") to compute invoices — summing invoice ids is meaningless for counting invoices.
- Option B uses count("billing_id") which counts rows; that can be incorrect in the presence of duplicate invoice IDs (if the same invoice appears multiple rows). The question explicitly asks for unique invoices, so we need a distinct count.
- Option C counts distinct patient_id which gives the number of unique patients per day, not unique invoices.
- Option D correctly sums amount_billed and uses count_distinct("billing_id") (in PySpark use F.countDistinct("billing_id") or F.countDistinct(col("billing_id"))) to get the unique invoice count.
Example expected output from the sample data:
- 2024-03-01: total_revenue = 1500 + 6500 = 8000, total_invoices = 2 (billing_id 401 and 403)
- 2024-03-02: total_revenue = 3000, total_invoices = 1
- 2024-03-03: total_revenue = 500, total_invoices = 1
Recommended PySpark code: from pyspark.sql import functions as F
daily_revenue_df = billing_df.groupBy("billing_date").agg( F.sum("amount_billed").alias("total_revenue"), F.countDistinct("billing_id").alias("total_invoices") )
Section: ELT With Spark SQL and Python
Comments
Loading comments...