A light weight library to connect, introspect, and query popular databases over a unified interface. This gem is intended for analtyical workloads. The library also provides database specific translations for common functions like date_trunc, date_add etc. The function tranlation is not comprehensive. But, it does provides good coverage for date handling, and some array handling as well.
Note
This is not an ORM nor will it cast types to ruby unless the underlying client does it out of the box. The goal here is to create an Architecture where new databases can be onboarded quickly.
Libraries like Sequel are amazing and comprehensive. However, its broad coverage also makes it more laborious to add new databases. Especially, ones with only HTTP endpoints for Ruby. We seem to be having an explosion of databases recently and a light weight interface will allow us to integrate faster.
The adapter only has 5 core methods (6 including the connection method). A YAML settings controls how it interacts with a particular db. It is relatively fast to add a new db. See the Druid implementation as an example. And here is its corresponding YAML settings file.
- Unified Interface: Connect to multiple database types using the same API
- SQL Function Translation: Automatically translates common SQL functions to database-specific syntax
- Connection Pooling: Built-in connection pool management for high-performance applications
- Rich Metadata: Extract table schemas, column information, and statistics
- Snowflake - High performance cloud warehouse
- Trino (formerly Presto) - Distributed SQL query engine
- Redshift - AWS data warehouse platform
- AWS Athena - AWS big data warehouse
- Apache Druid - Real-time analytics database
- DuckDB - In-process analytical database
- SQLite - Lightweight embedded database
- PostgreSQL - Full-featured RDBMS with advanced SQL support
- MySQL - Popular open-source database
- SQL Server - Microsoft's enterprise database
- ClickHouse - High performance analytical database
- Databricks - Lakehouse SQL warehouse
- Google BigQuery - Google Cloud serverless warehouse
- MotherDuck - Hosted DuckDB service
Install it yourself as:
gem install dwhrequire 'dwh'
# Connect to Druid
druid = DWH.create(:druid, {
host: 'localhost',
port: 8080,
protocol: 'http'
})
# basic query execution
results = druid.execute("SELECT * FROM web_sales", format: :csv)
# Connect to SQLite for local analytics
sqlite = DWH.create(:sqlite, {
file: 'path/to/analytics.db'
})
# Query with optimized WAL mode enabled by default
results = sqlite.execute("SELECT * FROM sales_data", format: :array)Standardized API across adapters:
- connection
- Creates a reusuable connection based on config hash passed in
- tables(schema: nil, catalog: nil)
- returns a list of tables from the default connection or from the specified schema and catalog
- metadata(table_name, schema: nil, catalog: nil)
- provides metadata about a table
- stats(table_name, date_column: nil)
- provides table row count and date range
- execute(sql, format: :array, retries: 0)
- runs a query and returns in given format
- execute_stream(sql, io, stats: nil)
- runs a query and streams it as csv into the given io
Certain databases have to be tested via docker. Those tests will try to launch docker compose services in test/support/compose*.yml
Run Unit Tests:
bundle exec rake test:unitRun tests on RDBMS dbs:
bundle exec rake test:system:rdbms Run tests on druid:
bundle exec rake test:system:druid Cloud warehouse tests (test/system/cloud_*_test.rb) run against real accounts and skip
unless configured through environment variables, so no credentials live in the repo:
| Adapter | Variables |
|---|---|
| Athena | ATHENA_S3_OUTPUT, optional ATHENA_REGION, ATHENA_DATABASE; AWS credentials from the environment |
| BigQuery | BIGQUERY_PROJECT, optional BIGQUERY_DATASET; GOOGLE_APPLICATION_CREDENTIALS or gcloud ADC. Creates its own fixture tables |
| Databricks | DATABRICKS_HOST, DATABRICKS_WAREHOUSE, DATABRICKS_CLIENT_ID, DATABRICKS_CLIENT_SECRET |
| Redshift | REDSHIFT_HOST, REDSHIFT_PASSWORD, optional REDSHIFT_USER, REDSHIFT_DATABASE, REDSHIFT_PORT |
| Snowflake | SNOWFLAKE_ACCOUNT, SNOWFLAKE_PAT, optional SNOWFLAKE_DATABASE; key-pair test needs SNOWFLAKE_USER, SNOWFLAKE_PRIVATE_KEY |
BIGQUERY_PROJECT=my-project BUNDLE_WITH=development bundle exec ruby -Itest test/system/cloud_bigquery_test.rbAfter checking out the repo, run bin/setup to install dependencies. Then, run rake test to run the tests. You can also run bin/console for an interactive prompt.
To install this gem onto your local machine, run bundle exec rake install.
Bug reports and pull requests are welcome on GitHub at https://github.com/stratasite/dwh.
This project is available as open source under the terms of the MIT License.
Current version: 0.6.3