error in Rails/UniqueValidationWithoutIndex

[seandilda]

---

Expected behavior

To not raise an exception

Actual behavior

Here's the backtrace generated by running rubocop -d <filename>:

Inspecting 1 file
Scanning /usr/src/app/app/models/broken_model.rb
An error occurred while Rails/UniqueValidationWithoutIndex cop was inspecting /usr/src/app/app/models/broken_model.rb:2:2.
undefined method `type' for nil:NilClass
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:82:in `each_content'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:62:in `each'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:62:in `map'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:62:in `build_columns'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:51:in `initialize'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:27:in `new'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:27:in `block in build!'

/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:37:in `block in each_table'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:34:in `each'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:34:in `each_table'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:26:in `build!'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader/schema.rb:12:in `initialize'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader.rb:34:in `new'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader.rb:34:in `load!'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/rails/schema_loader.rb:18:in `load'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/cop/rails/unique_validation_without_index.rb:128:in `schema'
/bundle/ruby/2.6.0/gems/rubocop-rails-2.5.0/lib/rubocop/cop/rails/unique_validation_without_index.rb:35:in `on_send'
/bundle/ruby/2.6.0/gems/rubocop-0.80.1/lib/rubocop/cop/commissioner.rb:57:in `block (2 levels) in trigger_responding_cops'

Steps to reproduce the problem

Here's the sample model I'm using:

class BrokenModel < ApplicationRecord
  validates :name, uniqueness: { scope: :second_attribute }
end

RuboCop version

Include the output of rubocop -V or bundle exec rubocop -V if using Bundler. Here's an example:

$ rubocop -V
0.80.1 (using Parser 2.7.0.5, running on ruby 2.6.5 x86_64-linux)

[koic]

Thanks for the feedback. Can you show me db/schema.rb?

[FraDim]

@koic: I get the same error. In my case, it is due to a table that only has a primary key (no other columns). Here is the relevant part of the db/schema.rb

create_table "data_migrations", primary_key: "version", id: :string, force: :cascade do |t|
end

[seandilda]

That particular model is not in the schema.

I originally hit this bug with a model called Building::Location and here its its schema entry:

  create_table "building_locations", id: :integer, options: "ENGINE=InnoDB DEFAULT CHARSET=utf8mb4", force: :cascade do |t|
    t.integer "building_id", null: false
    t.string "name", null: false
    t.integer "building_floor_plan_id"
    t.decimal "latitude", precision: 16, scale: 13
    t.decimal "longitude", precision: 16, scale: 13
    t.string "map_type"
    t.string "map_file_name"
    t.binary "map_data", limit: 16777215
    t.text "notes"
    t.datetime "created_at"
    t.datetime "updated_at"
    t.index ["building_floor_plan_id"], name: "index_building_locations_on_building_floor_plan_id"
    t.index ["building_id", "name"], name: "index_building_locations_on_building_id_and_name", unique: true
end

I also hit it on a model called device with the following validates line:

  validates :management_ip, format: { with: Resolv::IPv4::Regex }, uniqueness: true,
                            allow_nil: true, if: :management_ip_changed?

[FraDim]

Apologies @seandilda, I didn't mean to overtake this issue – just got overly excited that someone posted about the issue I was having, minutes before I went looking for it.

You seem to have a completely different bug than me.

[seandilda]

@FraDim no worries. Its still possible they're the same underlying issue.

[koic]

I reproduced the error and opened PR #218. Thank you!