ClickHouse SQLAlchemy

ClickHouse dialect for SQLAlchemy to ClickHouse database.

Installation

The package can be installed using pip:

pip install clickhouse-sqlalchemy

Interfaces support

• native [recommended] (TCP) via clickhouse-driver
• http via requests

Connection Parameters

ClickHouse SQLAlchemy uses the following syntax for the connection string:

'clickhouse+<driver>://<user>:<password>@<host>:<port>/<database>[?key=value..]'

Where:

• 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.

HTTP

• 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:

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

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

'clickhouse://user:[email protected]:8124/test?protocol=https'

Where 8124 is proxy port.

If you need control over the underlying HTTP connection, pass a requests.Session instance to create_engine(), like so:

from sqlalchemy import create_engine
from requests import Session

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

engine = create_engine(uri, connect_args={'http_session': Session()})

Native

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:

'clickhouse+native://default:@localhost/test'

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

Features

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)
    other_value = Column(
        types.DateTime,
        clickhouse_codec=('DoubleDelta', 'ZSTD'),
    )

    __table_args__ = (
        engines.Memory(),
    )

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

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__
table.create()
another_table.create()


another_table.drop()
table.drop()

Basic INSERT clause support

Simple batch INSERT:

from datetime import date, timedelta
from sqlalchemy import func

today = date.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(Rate.day)) \
    .filter(Rate.day > today - timedelta(20)) \
    .scalar()

session.query(Rate.value) \
    .order_by(Rate.day.desc()) \
    .first()

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

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

Advanced INSERT clause support

INSERT FROM SELECT statement:

from sqlalchemy import cast

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

# Emits single INSERT FROM SELECT statement
session.execute(
    another_table.insert()
    .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(
    Rate.day.label('date'),
    Rate.value.label('x')
)
select_another_rate = session.query(
    another_table.c.day.label('date'),
    another_table.c.value.label('x')
)

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]) \
    .all()

print(rv)

Supported ClickHouse-specific SQL

• SELECT query:

  • WITH TOTALS
  • 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}) \
    .scalar()

print(rv)

Running tests

mkvirtualenv testenv && python setup.py test

pip will automatically install all required modules for testing.

License

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.