Downloads

Blog

Column Store Database Benchmarks: MariaDB ColumnStore vs. Clickhouse vs. Apache Spark

March 17, 2017

Author

Alexander Rubin

Insight for DBAs

MySQL

Share this Post:

MariaDB no longer meeting your needs?

Migrate to Percona software for MySQL – an open source, production-ready, and enterprise-grade MySQL alternative.

Learn More

This blog shares some column store database benchmark results and compares the query performance of MariaDB ColumnStore v.1.0.7 (based on InfiniDB), ClickHouse, and Apache Spark.

I’ve already written about ClickHouse.

The purpose of the benchmark is to evaluate how these three solutions perform on a single large server with many CPU cores and large amounts of RAM. All systems are massively parallel (MPP) databases, designed to utilize many cores for SELECT queries.

Datasets

Wikipedia page counts (~26 billion rows for 2008)
Query analytics data from Percona Monitoring and Management
Online shop orders

This post focuses on Wikipedia page counts. Other datasets will be covered separately.

Databases, Versions, and Storage Engines

MariaDB ColumnStore v1.0.7 (ColumnStore engine)
Yandex ClickHouse v1.1.54164 (MergeTree engine)
Apache Spark v2.1.0 (Parquet and ORC)

All tests were run on a single server.

Hardware

CPU: 2 physical, 32 cores (64 threads)
RAM: 256 GB
Disk: Samsung SSD 960 PRO 1TB (NVMe)

Data Sizes

Dataset	ColumnStore	ClickHouse	MySQL	Spark (Parquet)	Spark (ORC)
Wikistat	374.24 GB	211.3 GB	n/a (>2 TB)	395 GB	273 GB
Query metrics	61.23 GB	28.35 GB	520 GB
Store Orders	9.3 GB	4.01 GB	46.55 GB

Query Performance

Warm Cache

Query	Spark	ClickHouse	ColumnStore
count(*)	5.37	2.14	30.77
group by month	205.75	16.36	259.09
top 100 pages	750.35	171.22	1640.7

Cold Cache

Query	Spark	ClickHouse	ColumnStore
count(*)	21.93	8.01	139.01
group by month	217.88	16.65	420.77
top 100 pages	887.43	182.56	1703.19

Partitioning and Primary Keys

ClickHouse uses primary keys to scan only relevant data chunks:

SELECT count(*), toMonth(date) AS mon
FROM wikistat
WHERE toYear(date)=2008 AND toMonth(date)=1
GROUP BY mon;

SELECT count(*), toMonth(date) AS mon

FROM wikistat

WHERE toYear(date)=2008 AND toMonth(date)=1

GROUP BY mon;

ColumnStore requires rewriting queries using date ranges for partition elimination:

SELECT count(*), month(date) AS mon
FROM wikistat
WHERE date BETWEEN '2008-01-01' AND '2008-01-31'
GROUP BY mon;

SELECT count(*), month(date) AS mon

FROM wikistat

WHERE date BETWEEN '2008-01-01' AND '2008-01-31'

GROUP BY mon;

Working with Large Datasets

Large GROUP BY operations require significant memory due to hash table usage:

SELECT path, count(*), sum(hits)
FROM wikistat
GROUP BY path
ORDER BY sum_hits DESC
LIMIT 100;

SELECT path, count(*), sum(hits)

FROM wikistat

GROUP BY path

ORDER BY sum_hits DESC

LIMIT 100;

ColumnStore does not support disk spill for GROUP BY, so memory tuning may be required:

<NumBlocksPct>10</NumBlocksPct>

1	<NumBlocksPct>10</NumBlocksPct>

SQL Support

Feature	Spark	ClickHouse	ColumnStore
INSERT	Yes	Yes	Yes
UPDATE	No	No	Yes
DELETE	No	No	Yes
Window functions	Yes	No	Yes

Comparison

Solution	Advantages	Disadvantages
MariaDB ColumnStore	MySQL compatibility Supports UPDATE/DELETE Window functions	Slower SELECT queries No GROUP BY disk spill No MySQL replication
ClickHouse	Fastest performance Better compression Disk-based GROUP BY	No MySQL protocol
Apache Spark	Flexible storage ML integration	Slower queries No MySQL protocol

Conclusion

ClickHouse is the clear winner in this benchmark, showing significantly better performance and compression.

However, ColumnStore provides a MySQL-compatible interface, making it a strong option for migrations from MySQL.

Table Structure

CREATE TABLE wikistat (
  date date,
  time datetime,
  project varchar(20),
  subproject varchar(2),
  path varchar(1024),
  hits bigint,
  size bigint
) ENGINE=Columnstore;

CREATE TABLE wikistat (

date date,

time datetime,

project varchar(20),

subproject varchar(2),

path varchar(1024),

hits bigint,

size bigint

) ENGINE=Columnstore;

Queries

select count(*) from wikistat;

1	select count(*) from wikistat;

select count(*), month(date) as mon
from wikistat
where date between '2008-01-01' and '2008-10-31'
group by mon;

select count(*), month(date) as mon

from wikistat

where date between '2008-01-01' and '2008-10-31'

group by mon;

SELECT path, count(*), sum(hits)
FROM wikistat
GROUP BY path
ORDER BY sum_hits DESC
LIMIT 100;

SELECT path, count(*), sum(hits)

FROM wikistat

GROUP BY path

ORDER BY sum_hits DESC

LIMIT 100;

0 Comments

Oldest

Newest Most Voted

Inline Feedbacks

View all comments

Resources

Blog

MySQL

April 17, 2026

MariaDB’s Snapshot Isolation: A Fix That Breaks More Than It Fixes

Blog

MySQL

April 13, 2026

Auditing Login Attempts in MySQL and MariaDB

Blog

MongoDB

April 7, 2026

Percona ClusterSync for MongoDB 0.8.0: Up to 18x Faster Change Replication

Far
Enough.

Said no pioneer ever.

Get Started

Open source database software from experts who stand with you in production. Forever free from lock-in and other corporate BS.

Connect

Privacy

Legal

Security Center

MySQL, PostgreSQL, InnoDB, MariaDB, MongoDB and Kubernetes are trademarks for their respective owners.

Column Store Database Benchmarks: MariaDB ColumnStore vs. Clickhouse vs. Apache Spark

MariaDB no longer meeting your needs?

Datasets

Databases, Versions, and Storage Engines

Hardware

Data Sizes

Query Performance

Warm Cache

Cold Cache

Partitioning and Primary Keys

Working with Large Datasets

SQL Support

Comparison

Conclusion

Table Structure

Queries

MariaDB’s Snapshot Isolation: A Fix That Breaks More Than It Fixes

Auditing Login Attempts in MySQL and MariaDB

Percona ClusterSync for MongoDB 0.8.0: Up to 18x Faster Change Replication

Far Enough.

Far
Enough.