Question

A retail company is building a data lake on Google Cloud. They ingest raw JSON and CSV files from hundreds of point-of-sale systems into a landing area. Requirements: files must be immediately queryable by ad hoc BigQuery jobs during the first 30 days, retained for at least 7 years for compliance, and stored as cheaply as possible after the active analysis window since they are rarely accessed afterward. The team wants a single, cost-optimized storage approach that avoids manual file movement. Which design best meets these needs?

Accepted Answer

Land files in a Cloud Storage bucket in the Standard class and configure a lifecycle rule to transition objects to Coldline after 30 days and keep them for 7 years; query with BigQuery external tables.

Answer

Load all files directly into a BigQuery native table with a 7-year partition expiration and delete the source files immediately after ingestion.

Answer

Store files in Bigtable to allow both low-latency queries and long-term retention with a single storage system.

Answer

Land files in a Cloud Storage Nearline bucket and manually move them to Archive storage using a scheduled Cloud Function after 30 days.