Neura Market
hydradatabase/columnar logo

hydradatabase/columnar

Free

Postgres-native columnar storage extension

FreeFree tier
Type
Open Source
Company
Hydra
Visit Website GitHub (38 stars)

About hydradatabase/columnar

Hydra Columnar is an open source, column-oriented Postgres extension that accelerates analytical workloads. It uses columnar storage, query parallelization, vectorized execution, and column-level caching to deliver fast query performance for aggregates (COUNT, SUM, AVG), WHERE clauses, and bulk operations. Hydra is a drop-in replacement for the standard Postgres Docker image and works with existing PostgreSQL tools and SQL. It operates as an extension via the table access method API (tableam) added in Postgres 12, not as a fork. Currently in alpha/beta stages, it is designed for use cases like clickstream analysis, web analytics, machine-generated data, structured logs, and events data. Limitations include no support for logical replication and only btree/hash indexes.

Key Features

Columnar storage for analytical queries
Query parallelization across CPU cores
Vectorized execution for faster data processing
Column-level caching to reduce I/O
Drop-in replacement for the standard Postgres Docker image
Uses PostgreSQL table access method API (tableam)
Optimized for aggregate functions (COUNT, SUM, AVG) and WHERE clauses
Supports bulk INSERTs, UPDATEs, and DELETEs
Open source under AGPL 3.0
Supports btree and hash indexes (no other index types)

Pros & Cons

Pros
  • Significantly faster analytical queries compared to row-oriented Postgres
  • Easy to deploy with provided Docker image as a drop-in replacement
  • Uses standard PostgreSQL SQL and ecosystem tools
  • Open source and free to use (AGPL 3.0)
  • Extends Postgres without forking, staying close to upstream
Cons
  • Currently in alpha/beta stages, not fully production-ready for all use cases
  • Limited index support: only btree and hash indexes, no GIN, GiST, or SP-GiST
  • Does not support logical replication for columnar tables
  • Not suitable for frequent large updates or small OLTP transactions
  • Some Postgres features may be unsupported (e.g., certain constraints)

Best For

Clickstream and traffic analysisWeb analyticsMachine-generated data processingStructured log analysisEvents data analysisReal-time analytics on large datasets

FAQ

Why is Hydra so fast?
Hydra achieves speed through columnar storage, query parallelization, vectorized execution, column-level caching, and tuning of PostgreSQL.
How do I start using the columnar format on Postgres?
Data is loaded into columnar format by default. You use Postgres normally; no special syntax required.
What operations is Hydra meant for?
Hydra is optimized for aggregates (COUNT, SUM, AVG), WHERE clauses, bulk INSERTs, UPDATEs, and DELETEs.
What is columnar not meant for?
Frequent large updates and small transactions are not ideal for columnar storage.
What Postgres features are unsupported on columnar?
Logical replication is unsupported. Columnar tables only support btree and hash indexes, and their associated constraints.
Is Hydra a fork of PostgreSQL?
No, Hydra is a PostgreSQL extension, not a fork. It uses the table access method API (tableam) introduced in Postgres 12.