How to pass arguments into a Rake task with environment in Rails? [duplicate]

Question

Just to follow up on this old topic; here's what I think a current Rakefile (since a long ago) should do there. It's an upgraded and bugfixed version of the current winning answer (hgimenez):

desc "Testing environment and variables"
task :hello, [:message]  => :environment  do |t, args|
  args.with_defaults(:message => "Thanks for logging on")
  puts "Hello #{User.first.name}. #{args.message}"   # Q&A above had a typo here : #{:message}
end

This is how you invoke it (http://guides.rubyonrails.org/v4.2/command_line.html#rake):

  rake "hello[World]" 

For multiple arguments, just add their keywords in the array of the task declaration (task :hello, [:a,:b,:c]...), and pass them comma separated:

  rake "hello[Earth,Mars,Sun,Pluto]" 

Note: the number of arguments is not checked, so the odd planet is left out:)

TLDR;

task :t, [args] => [deps] 

Original Answer

When you pass in arguments to rake tasks, you can require the environment using the :needs option. For example:

desc "Testing environment and variables"
task :hello, :message, :needs => :environment do |t, args|
  args.with_defaults(:message => "Thanks for logging on")
  puts "Hello #{User.first.name}. #{args.message}"
end

Updated per @Peiniau's comment below

As for Rails > 3.1

task :t, arg, :needs => [deps] # deprecated

Please use

task :t, [args] => [deps] 

Just for completeness, here the example from the docs mentioned above:

   task :name, [:first_name, :last_name] => [:pre_name] do |t, args|
     args.with_defaults(:first_name => "John", :last_name => "Dough")
     puts "First name is #{args.first_name}"
     puts "Last  name is #{args.last_name}"
   end

Notes:

• You may omit the #with_defaults call, obviously.
• You have to use an Array for your arguments, even if there is only one.
• The prerequisites do not need to be an Array.
• args is an instance of Rake::TaskArguments.
• t is an instance of Rake::Task.

An alternate way to go about this: use OS environment variables. Benefits of this approach:

• All dependent rake tasks get the options.
• The syntax is a lot simpler, not depending on the rake DSL which is hard to figure out and changes over time.

I have a rake task which requires three command-line options. Here's how I invoke it:

$ rake eaternet:import country=us region=or agency=multco

That's very clean, simple, and just bash syntax, which I like. Here's my rake task. Also very clean and no magic:

task import: [:environment] do
  agency = agency_to_import
  puts "Importing data for #{agency}..."
  agency.import_businesses
end

def agency_to_import
  country_code = ENV['country'] or raise "No country specified"
  region_slug  = ENV['region']  or raise "No region specified"
  agency_slug  = ENV['agency']  or raise "No agency specified"
  Agency.from_slugs(country_code, region_slug, agency_slug)
end

This particular example doesn't show the use of dependencies. But if the :import task did depend on others, they'd also have access to these options. But using the normal rake options method, they wouldn't.

While these solutions work, in my opinion this is overly complicated.

Also, if you do it this way in zsh, you'll get errors if the brackets in your array aren't escaped with '\'.

I recommend using the ARGV array, which works fine, is much simpler, and is less prone to error. E.g:

namespace :my_example do
  desc "Something"
  task :my_task => :environment do
    puts ARGV.inspect
  end
end

then

rake my_example:my_task 1 2 3

#=>  ["my_example:my_task", "1", "2", "3"]

The only thing you need to keep in mind is that ARGV[0] is the process name, so use only ARGV[1..-1].

I realize that strictly speaking this does not answer the question, as it does not make use of :environment as part of the solution. But OP did not state why he included that stipulation so it might still apply to his use case.