Question

A retail analytics team stores several terabytes of Parquet and CSV files in an Azure Data Lake Storage Gen2 account. Data analysts need to run occasional ad-hoc SQL queries directly against these files without loading the data into a database first. Query frequency is irregular—sometimes several times a day, sometimes nothing for a week. The team wants to minimize cost by paying only for the data processed by each query and avoid managing any provisioned infrastructure. Which solution should you recommend?

Accepted Answer

Use the serverless SQL pool in Azure Synapse Analytics to query the files in place using OPENROWSET

Answer

Provision a dedicated SQL pool in Azure Synapse Analytics and load the files with COPY INTO before querying

Answer

Deploy an Azure SQL Managed Instance and use PolyBase external tables to reach the data lake

Answer

Create an Azure Databricks all-purpose cluster and keep it running to serve interactive queries