SQL
Definition
Structured Query Language - standard language for managing and querying relational databases, essential for data manipulation and retrieval.
Use Cases
- Netflix: Business intelligence and operational reporting on large-scale datasets — Netflix has publicly described using SQL-based analytics with Presto (now Trino) to query data stored in a data lake (e.g., on Amazon S3) for interactive analysis and reporting. (Enabled analysts and engineers to run ad-hoc queries on large datasets and support data-driven decision-making with faster access to insights.)
- Airbnb: Company-wide analytics and experimentation analysis — Airbnb has publicly discussed using SQL as a primary interface for analytics (e.g., querying data in warehouses/lakes) and building internal tooling that allows employees to explore and analyze data using SQL. (Improved self-service analytics, helping teams measure experiments, monitor key metrics, and make product decisions based on data.)
- Uber: Analytics on event and operational data for performance monitoring and business reporting — Uber has shared using SQL-based query engines (commonly Presto/Trino and similar systems) to query large datasets for analytics workloads across the organization. (Supported scalable, interactive analytics for many teams, improving visibility into operations and accelerating investigation of business and system performance.)
Frequently Asked Questions
- What’s the difference between SQL and NoSQL?
- SQL is a language used mainly with relational databases, where data is organized into tables with rows and columns and relationships can be enforced (like customers and orders). NoSQL is a category of databases that often store data in formats like documents, key-values, or wide columns, and typically prioritize flexible schemas and horizontal scaling. You can still query NoSQL databases, but the query language and data model usually differ from traditional SQL tables.
- When should I use SQL?
- Use SQL when your data fits well into structured tables and you need reliable querying, reporting, and joins across related data (e.g., customers, orders, invoices). SQL is a strong choice for transactional systems (OLTP) like e-commerce checkouts, as well as analytics and reporting (OLAP) in data warehouses. If your data is highly unstructured, changes shape frequently, or you need extremely low-latency key-value access at massive scale, a NoSQL option may be a better fit.
- How much does SQL cost?
- SQL itself is a language standard and is free to use. Costs come from the database system and where it runs: (1) database licensing (some engines are open source like PostgreSQL/MySQL; others are commercial like Microsoft SQL Server or Oracle Database), (2) infrastructure (compute, storage, backups), and (3) managed service pricing (instance size, storage type, I/O, high availability, read replicas, and data transfer). In cloud managed databases, your bill is typically driven by the database instance/compute tier, storage consumed, and any additional features like multi-zone availability or serverless scaling.
Category: data
Difficulty: intermediate
Related Terms
See Also