Skip to content
ClickHouse dialect for SQLAlchemy
Branch: master
Clone or download
Latest commit c0d2952 Mar 5, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
clickhouse_sqlalchemy Fix wrong encoding error Mar 2, 2019
.travis.yml Numeric type support fixes Jan 26, 2019 Version 0.0.10 changelog Feb 24, 2019
LICENSE Initial commit Apr 1, 2017
README.rst Add how to contribute section in README.rst Sep 4, 2018


ClickHouse SQLAlchemy

ClickHouse dialect for SQLAlchemy to ClickHouse database.


The package can be installed using pip:

pip install clickhouse-sqlalchemy

Interfaces support

Connection Parameters

ClickHouse SQLAlchemy uses the following syntax for the connection string:



  • driver is driver to use. Possible choices: http, native. http is default.
  • database is database connect to. Default is default.

Drivers options

There are several options can be specified in query string.


  • port is port ClickHouse server is bound to. Default is 8123.
  • timeout in seconds. There is no timeout by default.
  • protocol to use. Possible choices: http, https. http is default.

Connection string to database test in default ClickHouse installation:


When you are using nginx as proxy server for ClickHouse server connection string might look like:


Where 8124 is proxy port.


Please note that native connection is not encrypted. All data including user/password is transferred in plain text. You should use this connection over SSH or VPN (for example) while communicating over untrusted network.

Connection string to database test in default ClickHouse installation:


All connection string parameters are proxied to clickhouse-driver. See it's parameters.


SQLAlchemy declarative support

Both declarative and constructor-style tables support:

from sqlalchemy import create_engine, Column, MetaData, literal

from clickhouse_sqlalchemy import Table, make_session, get_declarative_base, types, engines

uri = 'clickhouse://default:@localhost/test'

engine = create_engine(uri)
session = make_session(engine)
metadata = MetaData(bind=engine)

Base = get_declarative_base(metadata=metadata)

class Rate(Base):
    day = Column(types.Date, primary_key=True)
    value = Column(types.Int32)

    __table_args__ = (

another_table = Table('another_rate', metadata,
    Column('day', types.Date, primary_key=True),
    Column('value', types.Int32, server_default=literal(1)),

Tables created in declarative way have lowercase with words separated by underscores naming convention. But you can easy set you own via SQLAlchemy __tablename__ attribute.

Basic DDL support

You can emit simple DDL. Example CREATE/DROP table:

table = Rate.__table__


Basic INSERT clause support

Simple batch INSERT:

from datetime import date, timedelta
from sqlalchemy import func

today =
rates = [{'day': today - timedelta(i), 'value': 200 - i} for i in range(100)]

# Emits single INSERT statement.
session.execute(table.insert(), rates)

Common SQLAlchemy query method chaining

order_by, filter, limit, offset, etc. are supported:

session.query(func.count( \
    .filter( > today - timedelta(20)) \

session.query(Rate.value) \
    .order_by( \

session.query(Rate.value) \
    .order_by( \
    .limit(10) \

session.query(func.sum(Rate.value)) \

Advanced INSERT clause support


from sqlalchemy import cast

# Labels must be present.
select_query = session.query('day'),
    cast(Rate.value * 1.5, types.Int32).label('value')

# Emits single INSERT FROM SELECT statement
    .from_select(['day', 'value'], select_query)

Many but not all of SQLAlchemy features are supported out of the box.

UNION ALL example:

from sqlalchemy import union_all

select_rate = session.query('date'),
select_another_rate = session.query('date'),

union_all(select_rate, select_another_rate).execute().fetchone()

External data for query processing

Currently can be used with native interface.

ext = Table(
    'ext', metadata, Column('x', types.Int32),
    clickhouse_data=[(101, ), (103, ), (105, )], extend_existing=True

rv = session.query(Rate) \
    .filter(Rate.value.in_(session.query(ext.c.x))) \
    .execution_options(external_tables=[ext]) \


Supported ClickHouse-specific SQL

  • SELECT query:
    • SAMPLE
    • lambda functions: x -> expr
    • JOIN

See tests for examples.

Overriding default query settings

Set lower priority to query and limit max number threads to execute the request.

rv = session.query(func.sum(Rate.value)) \
    .execution_options(settings={'max_threads': 2, 'priority': 10}) \


Running tests

mkvirtualenv testenv && python test

pip will automatically install all required modules for testing.


ClickHouse SQLAlchemy is distributed under the MIT license.

How to Contribute

  1. Check for open issues or open a fresh issue to start a discussion around a feature idea or a bug.
  2. Fork the repository on GitHub to start making your changes to the master branch (or branch off of it).
  3. Write a test which shows that the bug was fixed or that the feature works as expected.
  4. Send a pull request and bug the maintainer until it gets merged and published.
You can’t perform that action at this time.