The database matters as much as the language — pick the wrong one and everything from performance to how much you pay gets harder. Ranked by real reported usage share (2026 developer survey data), grouped by category.
Learn SQL and then PostgreSQL specifically. Postgres is used by over half of developers surveyed, it's free and open-source, and the SQL you learn on it transfers almost directly to MySQL, SQL Server, and SQLite — the whole relational-database world speaks the same core language.
Relational (SQL) Databases
🐘
#1 · 55.6%
PostgreSQL
A free, open-source relational database known for strict standards-compliance, powerful advanced features (JSON columns, full-text search, custom types), and being the default choice for most new backend projects today.
🐬
#2 · 40.5%
MySQL
The database behind WordPress, most of the early web, and huge amounts of PHP-stack software — free and open-source, slightly simpler than Postgres, with an enormous amount of legacy hosting support.
📦
#3 · 37.5%
SQLite
Not a server at all — a full relational database stored as a single local file, with zero setup or configuration. Ships inside nearly every mobile app, browser, and desktop app that needs local storage.
🪟
#4 · 30.1%
Microsoft SQL Server
Microsoft's enterprise relational database — deeply integrated with the .NET/Windows ecosystem, common in large corporate IT environments that standardized on Microsoft's stack.
🦭
#7 · 22.5%
MariaDB
A community-driven fork of MySQL, created after Oracle acquired MySQL — nearly drop-in compatible with it, but fully open-source with no corporate ownership concerns.
🔴
#9 · 10.6%
Oracle Database
A powerful, expensive, heavily-licensed enterprise database — dominant in large legacy corporate and government systems (banking, insurance, government records) that were built decades ago and never migrated off it.
☕
#14 (tie) · 5%
H2 Database
A lightweight, embeddable Java SQL database — extremely common as the test/dev database swapped in for a "real" database (like Postgres) during automated testing in Java/Spring projects.
🗃️
#16 · 4.8%
Microsoft Access
A desktop database bundled with Microsoft Office — built for small, single-office applications (inventory trackers, simple internal tools) built by non-specialist office staff, not for real production software.
🔑 Key-Value & Cache
⚡
#5 · 28%
Redis
An in-memory key-value store used for caching, session storage, and real-time features (leaderboards, rate limiting, pub/sub messaging) — blazing fast because it lives in RAM, almost always paired alongside a "real" database rather than replacing one.
📊
#10 · 9.8%
DynamoDB
Amazon's fully-managed, serverless key-value/document database — scales automatically with no server to manage, deeply tied into the AWS ecosystem.
🔷
#24 · 2.4%
Valkey
A Linux Foundation-backed open-source fork of Redis, created after Redis changed its license terms — nearly identical to Redis in behavior, for developers who want to stay fully open-source.
📄 Document & NoSQL
🍃
#6 · 24%
MongoDB
The best-known document database — stores flexible, JSON-like records instead of rigid rows and columns, popular in JavaScript-stack apps (the "M" in the classic MERN stack) where the schema changes often during early development.
🔥
#13 · 5.7%
Cloud Firestore
Google Firebase's document database — real-time syncing to connected clients out of the box, tightly integrated with the rest of the Firebase platform (auth, hosting, functions).
🔥
#14 (tie) · 5%
Firebase Realtime Database
Firebase's original real-time database, predating Firestore — a single large JSON tree synced live to every connected client, simpler than Firestore but less structured for complex queries.
🐦
#22 · 2.9%
Cassandra
A wide-column database built for massive write-heavy workloads spread across many servers with no single point of failure — originally built at Facebook, now widely used anywhere data volume outgrows a single-server database.
📌
#28 (tie) · 1%
PocketBase
A single-file, open-source backend built on SQLite — bundles a database, auth, and file storage into one small self-hosted executable, popular for indie/solo projects that want a Firebase-like experience without the cloud bill.
☁️ Cloud, Analytics & Backend-as-a-Service
📈
#11 · 6.5%
Google BigQuery
Google Cloud's fully-managed data warehouse — built to run massive analytical SQL queries across billions of rows in seconds, priced by how much data you scan rather than a server you keep running.
⚡
#12 · 6%
Supabase
An open-source "Firebase alternative" built on real PostgreSQL — gives you a hosted Postgres database plus auth, file storage, and auto-generated APIs, aimed at developers who want Firebase's convenience with SQL underneath.
🌐
#17 · 4.6%
Cosmos DB
Microsoft Azure's globally-distributed, multi-model database — can behave like a document store, key-value store, or graph database depending on the API you choose, built for apps that need to run in multiple regions worldwide.
❄️
#18 · 4.1%
Snowflake
A cloud data warehouse built around separating storage from compute — you can scale query power up and down independently of how much data you store, popular for large enterprise analytics teams.
🧱
#20 · 3.4%
Databricks SQL
SQL querying built on top of the Databricks "lakehouse" platform, blending data-warehouse-style querying with the massive raw-file storage of a data lake — popular with data engineering and ML teams already using Databricks/Spark.
🦆
#21 · 3.3%
DuckDB
Sometimes called "SQLite for analytics" — an embedded, zero-setup database optimized for fast analytical queries over local files (CSV, Parquet), popular with data scientists working locally in Python/R notebooks.
🔺
#27 · 2.3%
Amazon Redshift
AWS's data warehouse, built on a modified version of PostgreSQL under the hood — designed for large-scale analytical queries across an organization's AWS-hosted data.
🪳
#28 (tie) · 1%
CockroachDB
A distributed SQL database designed to survive entire data-center outages without losing data — speaks standard SQL (Postgres-compatible wire protocol) while replicating data across regions automatically.
🔎 Search & Analytics Engines
🔍
#8 · 16.7%
Elasticsearch
A search and analytics engine built for fast full-text search and log analysis at scale — not a general-purpose database, but the standard choice whenever "search this text really fast" is the actual requirement.
📊
#24 (tie) · 2.4%
ClickHouse
An open-source columnar database built for extremely fast analytical queries over huge datasets — often used for real-time dashboards and analytics where traditional row-based databases get too slow.
🕸️ Specialized (Graph, Time-Series, Immutable)
🕸️
#23 · 2.6%
Neo4j
The best-known graph database — stores data as nodes and relationships rather than tables, built for problems that are naturally about connections (social networks, recommendation engines, fraud detection).
⏱️
#19 · 3.7%
InfluxDB
A database built specifically for time-series data — sensor readings, server metrics, IoT device data — where every record is timestamped and queries almost always involve a time range.
🔵
#26 · 2.4%
IBM Db2
IBM's long-running enterprise relational database — still common on mainframes and large legacy corporate systems (banking, insurance, government), similar niche to Oracle.
🧬
#30 · 0.6%
Datomic
An unusual immutable database — instead of overwriting data, every change is recorded as a new fact, so you can query the entire database as it existed at any point in its history. Built primarily for use with Clojure.
Usage percentages reflect a 2026 developer-tools survey; most developers use more than one database, so shares don't sum to 100%.