Migration library for Clickhouse
Project description
Clickhouse Migrator
Clickhouse is known for its scale to store and fetch large datasets.
Development and Maintenance of large-scale db systems many times requires constant changes to the actual DB system. Holding off the scripts to migrate these will be painful.
We found there is nothing existing earlier and developed one inspired by, Flyway, Alembic
This is a python library, which you can execute as a pre-hook using sys python. Or as a migration framework before deployment/server-startup in your application as required.
Publishing to pypi
- python -m build
- python -m twine upload --verbose --repository pypi dist/*
Installation
You can install from pypi using pip install clickhouse-migrator
.
Usage
# <= v1.0.4
from migration_lib.migrate import migrate
migrate(db_name, migrations_home, db_host, db_user, db_password, create_db_if_no_exists)
# > v1.0.4
from clickhouse_migrate.migrate import migrate
migrate(db_name, migrations_home, db_host, db_user, db_password, create_db_if_no_exists)
Parameter | Description | Default |
---|---|---|
db_name | Clickhouse database name | None |
migrations_home | Path to list of migration files | <project_root> |
db_host | Clickhouse database hostname | localhost |
db_password | ***** | **** |
create_db_if_no_exists | If the db_name is not present, enabling this will create the db |
True |
db_port | Database port incase your server runs on a non default port | None (defaults to 9000) |
queue_exec | Command Pipelining (wait for system mutations if any to complete) for every command | True |
Folder and Migration file patterns
The filenames are pretty similar to how flyway
keeps it.
Your first version filename should be prefixed with V1__
(double underscore)
These migrations are executed one by one, failures in between will stop and not further version files will be executed.
Multi statement and single statement migrations
If your migration is a single statement, you can create a file in the migration folder using the .sql extension and push your migration statement in there.
If you want to execute more than one statement in your migration, you can use a json file using the array syntax. Note that when using a json file, contents should be a valid json array as show. Ensure to keep migrations logical. Its not a good practise to push all migrations to one json file and neither is it wise to in all cases have them each statement in one file.
[
"CREATE TABLE pytest.sample1(id UInt32, name String) ENGINE MergeTree PARTITION BY tuple() ORDER BY tuple()",
"CREATE TABLE pytest.sample2(id UInt32, name String) ENGINE MergeTree PARTITION BY tuple() ORDER BY tuple()",
"CREATE TABLE pytest.sample3(id UInt32, name String) ENGINE MergeTree PARTITION BY tuple() ORDER BY tuple()"
]
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for clickhouse-migrator-1.0.6.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | e23f4e975774ffbc566ffcdc6350d675fe32910560ee0fb06b6009d040258aa0 |
|
MD5 | 04da898911830da34ce76bbcd70f2fb0 |
|
BLAKE2b-256 | 7385b4b4e13f7338f670d033a4cc52e17282dafcdb3c03994a70e293378f65ec |
Hashes for clickhouse_migrator-1.0.6-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | ebb77222597168307ea38643f58d8d5d3165878818342b15fb1603a4a4809f6d |
|
MD5 | 925453c2877bd535b81210599d8ae51c |
|
BLAKE2b-256 | 85abbfe093c865352846c04749e1d64acebcffbbd5804497a42cede87ae7e4a0 |