Skip to content
This repository

OData Consumer for Ruby

Octocat-spinner-32 config Modified test files to allow for more flexiblity in where the WCF ser… November 10, 2011
Octocat-spinner-32 features Revised error handling July 21, 2013
Octocat-spinner-32 gemfiles Updated 1.8.7 Gemfile to limit active support to v3 since 4 doesn't s… June 25, 2013
Octocat-spinner-32 lib Version bump to 0.1.6 March 26, 2014
Octocat-spinner-32 spec Merge commit '56c301e9c5511df447128cc3f193d7ce6abdb68d' into develop March 05, 2014
Octocat-spinner-32 test Updated README to show the new count method August 07, 2012
Octocat-spinner-32 .coveralls.yml Added coverage via coveralls June 21, 2013
Octocat-spinner-32 .gitignore Removed ruby-version and ruby-gemset from git July 03, 2013
Octocat-spinner-32 .rspec - Fixed a bug where if you had a WCF service with lowercase entities … August 24, 2011
Octocat-spinner-32 .simplecov Added simplecov for local coverage reporting and cleaned up Cucumber … June 23, 2013
Octocat-spinner-32 .travis.yml Removed 1.8.7 testing with Travis as celluloid needs >=1.9.3 November 16, 2013
Octocat-spinner-32 .yardopts Dumped RDoc and switched to markdown/YARD, added documentation, and r… August 07, 2012
Octocat-spinner-32 Version bump to 0.1.6 March 26, 2014
Octocat-spinner-32 Gemfile Removed Jeweler for gem creation and used bundler instead March 13, 2011
Octocat-spinner-32 Guardfile Modified Guardfile to run the correct spec when changing a class August 06, 2012
Octocat-spinner-32 LICENSE Changed license to MIT and added it to the gemspec November 19, 2013
Octocat-spinner-32 Added blog post reference for v0.1.0 February 21, 2013
Octocat-spinner-32 Rakefile Added simplecov for local coverage reporting and cleaned up Cucumber … June 23, 2013
Octocat-spinner-32 ruby_odata.gemspec Merge commit '56c301e9c5511df447128cc3f193d7ce6abdb68d' into develop March 05, 2014


The Open Data Protocol (OData) is a fantastic way to query and update data over standard Web technologies. The ruby_odata library acts as a consumer of OData services.

Build Status



You can install ruby_odata as a gem using:

gem install ruby_odata


Instantiating the Service

There are various options that you can pass when creating an instance of the service, these include:

  • username: username for http basic auth
  • password: password for http basic auth
  • verify_ssl: false if no verification, otherwise mode (OpenSSL::SSL::VERIFY_PEER is default)
  • rest_options: a hash of options that will be passed on to the rest-client calls. The passed in rest_options will be merged with the standard options that are set (username, password, and verify_ssl). This will allow you to set additional SSL settings. See the rest-client docs for more information. Note, the options that you pass in will take precedence over the previous 3 options, so it is possible to set/override the username, password, and verify_ssl options directly with this hash.
  • additional_params: a hash of query string params that will be passed on all calls (query, new, update, delete, batch)
  • namespace: a string based namespace to create your objects in. You can specify the namespace using periods as separators (like .NET, for example VisoftInc.Sample.Models) or using double colons as separators (like Ruby VisoftInc::Sample::Models). By providing a namespace you can prevent naming collisions in your applications.


When you point at a service, an AddTo method is created for you. This method takes in the new entity to create. To commit the change, you need to call the save_changes method on the service. To add a new category for example, you would simply do the following:

require 'ruby_odata'

svc = ""
new_category =
new_category.Name = "Sample Category"
category = svc.save_changes
puts category.to_json


To update an object, simply pass the modified object to the update_object method on the service. Updating, like adding requires you to call save_changes in order to persist the change. For example:

require 'ruby_odata'

svc = ""
new_category =
new_category.Name = "Sample Category"
category = svc.save_changes
puts category.to_json

category.Name = 'Updated Category'
result = svc.save_changes
puts "Was the category updated? #{result}"


Deleting an object involves passing the tracked object to the delete_object method on the service. Deleting is another function that involves the save_changes method (to commit the change back to the server). In this example, we'll add a category and then delete it.

require 'ruby_odata'

svc = ""
new_category =
new_category.Name = "Sample Category"
category = svc.save_changes
puts category.to_json

result = svc.save_changes
puts "Was the category deleted? #{result}"

Add Link

Adding a linkage between entities can now be performed outside of creation or modification of the objects. See the OData documents for more details. To add a link between entities, simply call the add_link method on the Service passing the parent object, the name of the navigation property, and the child object. Like all save operations, you need to call save_changes to persist the changes.

svc.add_link(<Parent>, <Navigation Property Name>, <Child>)


Querying is easy, for example to pull all the categories from the SampleService, you simply can run:

require 'ruby_odata'

svc = ""
categories = svc.execute
puts categories.to_json

You can also expand, add filters, order, skip records, and take only the top X records to the query before executing it. For example:


Expanding allows you to eagerly load other objects that are children of the root. You can use more than one expand on a query. For expanding grandchild and lower entities, you must pass in the full path from the root, for example Products.expand('Orders').expand('Orders/LineItems')

# Without expanding the query
prod1 = svc.execute
puts "Without expanding the query"
puts "#{prod1.to_json}\n"

# With expanding the query
prod1 = svc.execute
puts "With expanding the query"
puts "#{prod1.to_json}\n"

Lazy Loading

If you want to implement lazy loading, the ruby_odata Service allows you to perform this. You simply need to call the load_property method on the Service passing in the object and the navigation property to fill.

# Without expanding the query
prod1 = svc.execute.first
puts "#{prod1.to_json}\n"

# Use load_property for lazy loading
svc.load_property(prod1, "Category")
puts "#{prod1.to_json}\n"


The syntax for filtering can be found on the OData Protocol URI Conventions page. You can use more than one filter, if you call the filter method multiple times it will before an AND.

# You can access by ID (but that isn't is a filter)
# The syntax is just svc.ENTITYNAME(ID) which is shown in the expanding examples above

svc.Products.filter("Name eq 'Product 2'")
prod = svc.execute
puts "Filtering on Name eq 'Product 2'"
puts "#{prod.to_json}"

Note you can pass more than one filter in the string, for example (querying Netflix):

svc.Titles.filter("Rating eq 'PG' and ReleaseYear eq 1980")

Filters can also be chained, by doing this you will create an "and" filter (just like the last example) when it is passed to the server.

svc.Titles.filter("Rating eq 'PG'").filter("ReleaseYear eq 1980")

Combining Expanding and Filtering

The query operations follow a fluent interface, although they can be added by themselves as well as chained

svc.Products.filter("Name eq 'Product 2'").expand("Category")
prod = svc.execute
puts "Filtering on Name eq 'Product 2' and expanding"
puts "#{prod.to_json}"

Order By

You can order the results by properties of your choice, either ascending or descending. Order by are similar to expands in that you can use more than one of them on a query. For expanding grandchild and lower entities, you must pass in the full path from the root like would do on an expand

products = svc.execute

# Specifically requesting descending
svc.Products.order_by("Name desc")
products = svc.execute

# Specifically requesting ascending
svc.Products.order_by("Name asc")
products = svc.execute

Like the fiter method, order_by statements can also be chained like so:

svc.Products.order_by("Name asc").order_by("Price desc")


Skip allows you to skip a number of records when querying. This is often used for paging along with top.

products = svc.execute # => skips the first 5 items


Top allows you only retrieve the top X number of records when querying. This is often used for paging along with skip.
products = svc.execute # => returns only the first 5 items

Navigation Property Links Only Query

OData allows you to query navigation properties and only return the links for the entities (instead of the data). Note: You cannot use the links method and the count method in the same query

product_links = svc.execute # => returns URIs for the products under the Category with an ID of 1

Advanced Navigation Property Functions

There are instances where you may need to navigate down a level in order to form the proper query. Take for example Netflix's OData Service and their Genres Entity Collection, where you can access a Navigation Property (in this case Titles through the Genre and filter on it:'Horror%20Movies')/Titles?$filter=Name%20eq'Halloween'

In order to do this within ruby_odata, you can use the navigate method of the QueryBuilder to drill-down into the Navigation Property. This will allow you to perform filters, skips, orderbys, etc. against the children.

svc ="")
svc.Genres("'Horror Movies'").navigate("Titles").filter("Name eq 'Halloween'")
movies = svc.execute
movies.each { |m| puts m.Name }


Sometimes all you want to do is count records, for that, you can use the count method. This method can be combined with other options, such as filter but cannot be combined with the links method.

puts svc.execute # => 2

Partial feeds

OData allows services to do server-side paging in Atom by defining a next link. The default behavior is to repeatedly consume partial feeds until the result set is complete.

results = svc.execute # => retrieves all results in the Partials collection

If desired (e.g., because the result set is too large to fit in memory), explicit traversal of partial results can be requested via options:

svc = "", { :eager_partial => false }
results = svc.execute # => retrieves the first set of results returned by the server
if svc.partial? # => true if the last result set was a partial result set (i.e., had a next link)
  results.concat # => retrieves the next set of results
while svc.partial? # => to retrieve all partial result sets


Basic HTTP Authentication is supported via sending a username and password as service constructor arguments:

require 'ruby_odata'

svc = "", { :username => "bob", :password=> "12345" }

SSL/https Certificate Verification

The certificate verification mode can be passed in the options hash via the :verify_ssl key. For example, to ignore verification in order to use a self-signed certificate:

require 'ruby_odata'

svc = "", { :verify_ssl => false }

Or an OpenSSL integer constant can be passed as well:

require 'ruby_odata'

svc = "", { :verify_ssl => OpenSSL::SSL::VERIFY_PEER }

Default verification is OpenSSL::SSL::VERIFY_PEER. Note due to the way Ruby's Request object implements certificate checking, you CAN NOT pass OpenSSL::SSL::VERIFY_NONE, you must instead pass a boolean false.

Function Imports / Custom Service Methods

Function Imports are custom service methods exposed by the WCF Data Service. Each function import will be created as a method on the ruby_odata Service. When you make a call to one of these, they return a result immediately without the need to call execute or save_changes.


Instead of relying on looking at the EDMX directly, ruby_odata allows you to perform basic reflection on objects

Service Level Methods

  • Collections - You can look at the collections exposed by a service by accessing the collections method, which is a hash. The key is the name of the collection and the value is the hash with edmx_type, which returns the name of the type from the EDMX and :type, which is the local type that is created for you
  • Classes - To see the generated classes, you can utilize the classes method on the service. The return result is a hash where the key is the class name and the value is the class type.
  • Function Imports - You can find any function import (custom service methods) exposed by the service by accessing the function_imports method. This is a hash where the key is the Function Import name and the value is metadata about the Function Import.

Class Level Methods

  • Properties - You can call the class method properties on a generated class to see the method (properties) that were created. The returned result is a hash where the key is the property name and the value is metadata for the property like if it is nullable, the EDM Type, etc.


All of the tests are written using Cucumber going against a sample service (Found in test/RubyODataService/RubyODataService/*). The SampleService is an ASP.NET Web Site running a SQL Compact 4 Database, which gets generated at runtime, as well as the ADO.NET Entity Framework 4.1 and WCF Data Services October 2011 CTP. The reason for the CTP is that WCF Data Services currently shipping with .NET 4 doesn't support Entity Framework 4.1's "Code First" approach (e.g. no EDMX, all POCOs) In order to run the tests, you need to spin up IIS running a virtual directory of SampleService on port 8989 (http://localhost:8989/SampleService) and another instance running on port 44300.

NOTE The ports (8989 and 44300) and webserver (localhost by default) here are customizable thanks to /features/contants.rb. Take a look in there for the corresponding environment variables that you can set.

The SampleService requires IIS or IIS Express. IIS Express is a free download from Microsoft and the preferred approach to running the application. Once installed, there is a batch file found in /test called "iisExpress x64.bat" that will spin up the appropriate instances needed for the Cucumber tests. There is a also an "iisExpress x86.bat" file for those of you running a 32-bit machine. The only difference is the path to the Program Files directory. Once you run the batch file, the web server will spin up. To stop the server, use 'Q' and then enter or close the command window.

If you are having trouble with IIS Express, you may need to perform the following: Upon running the IIS Express installer copy the config folder from the IIS Express installed folder (e.g. c:\Progam Files (x86)\IIS Express\config) to the IIS folder in your home folder (e.g. c:\Users\Administrator\Documents\IISExpress). Within the newly copied config folder, copy the aspnet.config file from the templates\PersonalWebServer\aspnet.config folder into the config folder as well (e.g. c:\Users\Administrator\Documents\IISExpres\config\aspnet.config).

If you are testing on a Windows machine, you may encounter a problem with using Cucumber and Ruby 1.9.2. You will get a message when you fire up cucumber about missing msvcrt-ruby18.dll. The fix for this is to make sure you have the RubyInstaller DevKit installed, then do the following:

gem uninstall json
gem install json --platform=ruby -v 1.4.6

Once the SampleService is running, from the BASE ruby_odata directory, simply type rake, which will run the RSpec and Cucumber specs. You can also run them separately rake spec for RSpec and rake features for Cucumber.

Something went wrong with that request. Please try again.