Activecube: Multi-Dimensional Queries with Rails
Activecube is the library to make multi-dimensional queries to data warehouse, such as:
Cube.slice(
date: cube.dimensions[:date][:date].format('%Y-%m'),
currency: cube.dimensions[:currency][:symbol]
).measure(:count).
when(cube.selectors[:currency].in('USD','EUR')).to_sql
Cube, dimensions, metrics and selectors are defined in the Model, similary to
ActiveRecord.
Activecube uses Rails ActiveRecord in implementation.
In particular, you have to define all tables, used in
Activecube, as ActiveRecord tables.
Installation
Add this line to your application's Gemfile:
gem 'activecube'
And then execute:
$ bundle
Or install it yourself as:
$ gem install activecube
Usage
Basic steps to use ActiveCube are:
- Define your database schema in models, as you do with Rails
- Setup connection properties to data warehouse in config/database.yml. You can use multiple connections
if you use Rails 6 or higher
- Define cubes in models, sub-classed from Activecube::Base. Look
spec/models/test/transfers_cube.rb as example
- Make queries to the cubes
Check spec/cases/activecube_spec.rb for more examples.
Cube definition
Cube defined using the following attributes:
- table specifies, which physical database tables can be considered to query
table TransfersCurrency
table TransfersFrom
table TransfersTo
- dimension specifies classes used for slicing the cube
dimension date: Dimension::Date,
currency: Dimension::Currency
Or
dimension date: Dimension::Date
dimension currency: Dimension::Currency
- metric specifies which results expected from the cube queries
metric amount: Metric::Amount,
count: Metric::Count
Or
metric amount: Metric::Amount
metric count: Metric::Count
- selector is a set of expressions, which can filter results
selector currency: CurrencySelector,
transfer_from: TransferFromSelector,
transfer_to: TransferToSelector
Table definition
Tables are defined as regular active records, with additional optional attribute 'index':
index 'currency_id', cardinality: 4
which means that the table has an index onm currency_id field, with average number of different entries
of 10,000 ( 10^4). This creates a hint for optimizer to build queries.
Indexes can span multiple fields, as
index ['currency_id','date'], cardinality: 6
Note, that if you created combined index in database, you most probable will need to define all
indexed combinations, for example:
index ['currency_id'], cardinality: 4
index ['currency_id','date'], cardinality: 6
You can require using index in some cases. If required: true added, the table will be used only in case when this field is used
in query metric, dimension or selector.
index ['currency_id'], cardinality: 4, required: true
Query language
You use the cube class to create and execute queries.
Queries can be expressed as Arel query, SQL or executed against the database, returning results.
The methods used to contruct the query:
- slice defines which dimensions slices the results
- measure defines what to measure
- when defines which selectors to apply
- desc, asc, offset, limit are for ordering and limiting result set
After the query contructed, the following methods can be applied:
- to_sql to generate String SQL query from cube query
- to_query to generate Arel query
- query to execute query and return ResultSet
Managing Connections
You can control the connection used to construct and execute query by
ActiveRecord standard API:
ApplicationRecord.connected_to(database: :data_warehouse) do
cube = My::TransfersCube
cube.slice(
date: cube.dimensions[:date][:date].format('%Y-%m'),
currency: cube.dimensions[:currency][:symbol]
).measure(:count).query
end
will query using data_warehouse configuraton.
Alternatively you can use the method provided by activecube. It will
make the connection for the model or abstract class, which is super class for your models:
My::TransfersCube.connected_to(database: :data_warehouse) do |cube|
cube.slice(
date: cube.dimensions[:date][:date].format('%Y-%m'),
currency: cube.dimensions[:currency][:symbol]
).measure(:count).query
end
How it works
When you construct and execute cube query with any outcome ( sql, Arel query or ResultSet),
the same sequence of operations happen:
- Cube is collecting the query into a set of objects from the chain method call;
- Query is matched against the physical tables, the tables are selected that can serve the query or its part. For example, one table can provide one set of metrics, and the other can provide remaining;
- If possible, the variant is selected from all possible options, which uses indexes with the most cardinality
- Query is constructed using Arel SQL engine ( included in ActiveRecord ) using selected tables, and possibly joins
- If requested, the query is converted to sql ( using Arel visitor ) or executed with database connection
Optimization
The optimization on step #3 try to minimize the total cost of execution:
)
where
Optimization is done using the algorithm, which checks possible combinations of metrics and tables.
Development
After checking out the repo, run bin/setup
to install dependencies. Then, run rake spec
to run the tests. You can also run bin/console
for an interactive prompt that will allow you to experiment.
To install this gem onto your local machine, run bundle exec rake install
. To release a new version, update the version number in version.rb
, and then run bundle exec rake release
, which will create a git tag for the version, push git commits and tags, and push the .gem
file to rubygems.org.
RSPec tests
To run tests, you need clickhouse server installation.
Tests use database 'test' that have to be created in clickhouse as:
CREATE DATABASE test;
Check credentials for connection in spec/spec_helper.rb file.
By default clickhouse must reside on "clickhouse" server name, port 8123 with the default user access open.
Contributing
Bug reports and pull requests are welcome on GitHub at https://github.com/bitquery/activecube. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the Contributor Covenant code of conduct.
License
The gem is available as open source under the terms of the MIT License.
Code of Conduct
Everyone interacting in the Activecube project’s codebases, issue trackers, chat rooms and mailing lists is expected to follow the code of conduct.