Skip to main content

Versioning library for relational data

Project description

A library built on top of the SQLAlchemy ORM for versioning row changes to relational SQL tables.

Authors: Ryan Kirkman and Akshay Nanavati

Build Status

https://travis-ci.org/NerdWalletOSS/versionalchemy.svg?branch=master Documentation Status

Getting Started

$ pip install versionalchemy

Sample Usage

import sqlalchemy as sa
from sqlalchemy import create_engine
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.schema import UniqueConstraint

import versionalchemy as va
from versionalchemy.models import VAModelMixin, VALogMixin

MY_SQL_URL = '<insert mysql url here>'
engine = create_engine(MY_SQL_URL)
Base = declarative_base(bind=engine)

class Example(Base, VAModelMixin):
    __tablename__ = 'example'
    va_version_columns = ['id']
    id = sa.Column(sa.Integer, primary_key=True)
    value = sa.Column(sa.String(128))


class ExampleArchive(Base, VALogMixin):
    __tablename__ = 'example_archive'
    __table_args__ = (
        UniqueConstraint('id', 'va_version'),
    )
    id = sa.Column(sa.Integer)
    user_id = sa.Column(sa.Integer)

va.init()  # Only call this once
Example.register(ExampleArchive, engine)  # Call this once per engine, AFTER va.init

Latency

We used benchmark.py to benchmark the performance of versionalchemy. It times the performance of the SQLAlchemy core, ORM without VersionAclehmy and ORM with VersionAlchemy for n inserts (where n was variable). Some results are below.

n

Core Time

ORM Time

VA Time

10000

9.81 s

16.04 s

36.13

100000

98.78 s

158.87 s

350.84 s

VersionAlchemy performs roughly 2 times as bad as the ORM, which makes sense as we are doing roughly one additional insert per orm insert into the archive table.

Contributing

  • Make sure you have pip and virtualenv on your dev machine

  • Fork the repository and make the desired changes

  • Run make install to install all required dependencies

  • Run make lint tests to ensure the code is pep8 compliant and all tests pass. Note that the tests require 100% branch coverage to be considered passing

  • Open a pull request with a detailed explaination of the bug or feature

  • Respond to any comments. The PR will be merged if the travis CI build passes and the code changes are deemed sufficient by the admin

Style

  • Follow PEP8 with a line length of 100 characters

  • Prefer parenthesis to \ for line breaks

License

MIT License

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

versionalchemy-0.1.3.tar.gz (10.5 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page