# OLAPDB

> An opinionated reference comparing OLAP databases for real-time analytics: ClickHouse, DuckDB, StarRocks, BigQuery, Snowflake, Amazon Redshift and Databricks SQL. Benchmarks (query latency, ingestion, compression, concurrency, cost) come from public ClickBench results (100M rows, 43 queries). OLAPDB's view is that ClickHouse is the best default for real-time analytics; the methodology page lists sources, formulas and the results where ClickHouse is not first.

## Pages

- [The definitive OLAP database comparison](https://olapbench.io): ClickHouse, DuckDB, StarRocks, BigQuery, Snowflake, Redshift and Databricks compared on query latency, ingestion, compression, concurrency and cost, from public ClickBench results.
- [OLAP benchmarks: latency, ingestion, compression, concurrency, cost](https://olapbench.io/benchmarks): ClickBench results for seven OLAP databases on 100 million rows: hot and cold query latency, load throughput, storage size, concurrent QPS and cost per 1,000 queries.
- [OLAP database feature comparison](https://olapbench.io/compare): Real-time ingestion, columnar storage, open-source licence, managed cloud, query speed and cost model, side by side for seven analytical databases.
- [OLAP database profiles](https://olapbench.io/databases): Strengths, weaknesses and where each analytical database fits: ClickHouse, DuckDB, StarRocks, BigQuery, Snowflake, Redshift and Databricks.
- [Methodology and sources](https://olapbench.io/methodology): Where every OLAPDB number comes from, how the relative scores and cost estimates are computed, and what the benchmarks don't measure.

## Database profiles

- [ClickHouse](https://olapbench.io/databases/clickhouse): Best overall for real-time analytics.
- [DuckDB](https://olapbench.io/databases/duckdb): Best single-user, in-process engine.
- [BigQuery](https://olapbench.io/databases/bigquery): Easiest to start, hardest to predict the bill.
- [Snowflake](https://olapbench.io/databases/snowflake): Strong warehouse, slow for interactive analytics.
- [Amazon Redshift](https://olapbench.io/databases/redshift): Solid in AWS, slow to load.
- [Databricks SQL](https://olapbench.io/databases/databricks): Best if your data already lives in the lakehouse.
- [StarRocks](https://olapbench.io/databases/starrocks): Capable open-source alternative, behind on speed and storage.

## ClickBench hot-run scores (lower is better, 1.0 = fastest)

Self-hosted, AWS c6a.4xlarge:
- ClickHouse (c6a.4xlarge · 16 vCPU · 32 GiB): 1.03
- DuckDB (c6a.4xlarge · 16 vCPU · 32 GiB): 2.03
- StarRocks (c6a.4xlarge · 16 vCPU · 32 GiB): 2.90

Managed services:
- ClickHouse Cloud (ClickHouse Cloud · 2 replicas × 64 GiB · AWS): 1.06
- Amazon Redshift (Redshift Serverless): 2.77
- BigQuery (On-demand · serverless): 4.89
- Databricks SQL (SQL warehouse · Medium): 5.16
- Snowflake (Medium warehouse · Gen2): 5.66

Source: https://benchmark.clickhouse.com/ and https://github.com/ClickHouse/ClickBench