123456789_123456789_123456789_123456789_123456789_

Class: RuboCop::Cop::Style::CollectionQuerying

Relationships & Source Files
Super Chains via Extension / Inclusion / Inheritance
Class Chain:
self, ::RuboCop::Cop::AutoCorrector, ::RuboCop::Cop::Base, ::RuboCop::ExcludeLimit, NodePattern::Macros, RuboCop::AST::Sexp
Instance Chain:
self, ::RuboCop::Cop::RangeHelp, ::RuboCop::Cop::Base, ::RuboCop::Cop::AutocorrectLogic, ::RuboCop::Cop::IgnoredNode, ::RuboCop::Util, RuboCop::AST::Sexp
Inherits: RuboCop::Cop::Base
Defined in: lib/rubocop/cop/style/collection_querying.rb

Overview

Prefer Enumerable predicate methods over expressions with count.

The cop checks calls to count without arguments, or with a block. It doesn’t register offenses for count with a positional argument because its behavior differs from predicate methods (count matches the argument using ==, while any?, none? and one? use ===).

Note
 This cop doesn’t check length and size methods because they would yield false positives. For example, ::String implements length and size, but it doesn’t include Enumerable.

When count is used without a block, it measures the collection’s size, so the offense is corrected to empty? or !empty?. The predicate methods any?, none? and one? test element truthiness instead and would return different results for collections with falsey elements (e.g. [false].count.zero? is false, but [false].none? is true). For the same reason, count == 1 and count > 1 without a block are not flagged.

Examples:

# bad
x.count.positive?
x.count > 0
x.count != 0

# good
!x.empty?

# bad
x.count(&:foo?).positive?
x.count { |item| item.foo? }.positive?

# good
x.any?(&:foo?)
x.any? { |item| item.foo? }

# bad
x.count.zero?
x.count == 0

# good
x.empty?

# bad
x.count(&:foo?).zero?
x.count(&:foo?) == 0

# good
x.none?(&:foo?)

# bad
x.count(&:foo?) == 1

# good
x.one?(&:foo?)

AllCops:ActiveSupportExtensionsEnabled: false (default)

# good
x.count(&:foo?) > 1

AllCops:ActiveSupportExtensionsEnabled: true

# bad
x.count(&:foo?) > 1

# good
x.many?(&:foo?)

Cop Safety Information:

  • The cop is unsafe because receiver might not include Enumerable, or it has nonstandard implementation of count or any replacement methods.

    Blockless count is corrected to empty?, which Enumerable doesn’t define. Range, Enumerator (including lazy ones), Struct, StringIO and any class that just includes Enumerable respond to count but not to empty?, so (1…​1).count.zero? is true while (1…​1).empty? raises NoMethodError.

    Autocorrection is unsafe when replacement methods don’t iterate over every element in collection and the given block runs side effects:

    x.count(&:method_with_side_effects).positive?
# calls {method_with_side_effects} on every element

x.any?(&:method_with_side_effects)
# calls {method_with_side_effects} until first element returns a truthy value

Constant Summary

::RuboCop::Cop::Base - Inherited

DEPARTMENT_SEVERITIES, EMPTY_OFFENSES, RESTRICT_ON_SEND

::RuboCop::Cop::RangeHelp - Included

BYTE_ORDER_MARK, NOT_GIVEN

Class Attribute Summary

::RuboCop::Cop::AutoCorrector - Extended

support_autocorrect?

::RuboCop::Cop::Base - Inherited

.gem_requirements, .lint?,
.support_autocorrect?

Returns if class supports autocorrect.
.support_multiple_source?

Override if your cop should be called repeatedly for multiple investigations Between calls to on_new_investigation and on_investigation_end, the result of processed_source will remain constant.

Class Method Summary

::RuboCop::Cop::Base - Inherited

.autocorrect_incompatible_with

List of cops that should not try to autocorrect at the same time as this cop.
.badge

Naming.
.callbacks_needed, .cop_name, .department,
.documentation_url

Returns a url to view this cops documentation online.
.exclude_from_registry

Call for abstract Cop classes.
.inherited,
.joining_forces

Override and return the Force class(es) you need to join.
.match?

Returns true if the cop name or the cop namespace matches any of the given names.
.new,
.requires_gem

Register a version requirement for the given gem name.
.restrict_on_send

Reserved for Commissioner.

::RuboCop::ExcludeLimit - Extended

exclude_limit

Sets up a configuration option to have an exclude limit tracked.
transform

Instance Attribute Summary

::RuboCop::Cop::Base - Inherited

#active_support_extensions_enabled?, #always_autocorrect?, #config, #config_to_allow_offenses, #config_to_allow_offenses=, #contextual_autocorrect?, #disabled_autocorrect?,
#preview?

Whether the user opted in to unstable behavior, with --preview or AllCops: Preview.
#processed_source, #project_index, #string_literals_frozen_by_default?, #target_satisfies_all_gem_version_requirements?

::RuboCop::Cop::AutocorrectLogic - Included

#autocorrect?, #autocorrect_enabled?, #autocorrect_requested?, #autocorrect_with_disable_uncorrectable?, #correctable?, #disable_uncorrectable?, #safe_autocorrect?,
#skipped_unsafe_correction_with_disable_uncorrectable?

Whether a correction exists but is skipped because it is unsafe in a safe autocorrect run, while --disable-uncorrectable asks for a todo comment on everything this run cannot correct.
#contextual_autocorrect_held_back?

When LSP is enabled or the --editor-mode option is on, it is considered as editing source code, and autocorrection with AutoCorrect: contextual will not be performed.

Instance Method Summary

::RuboCop::Cop::RangeHelp - Included

#add_range,
#arguments_range

A range containing the first to the last argument of a method call or method definition.
#column_offset_between,
#contents_range

A range containing only the contents of a literal with delimiters (e.g.
#directions,
#effective_column

Returns the column attribute of the range, except if the range is on the first line and there’s a byte order mark at the beginning of that line, in which case 1 is subtracted from the column value.
#final_pos

rubocop:disable-next Metrics/ParameterLists.
#move_pos, #move_pos_str, #range_between, #range_by_whole_lines, #range_with_comments, #range_with_comments_and_lines, #range_with_surrounding_comma, #range_with_surrounding_space, #source_range

::RuboCop::Cop::Base - Inherited

#add_global_offense

Adds an offense that has no particular location.
#add_offense

Adds an offense on the specified range (or node with an expression) Unless that offense is disabled for this range, a corrector will be yielded to provide the cop the opportunity to autocorrect the offense.
#begin_investigation

Called before any investigation.
#callbacks_needed

 — grouped under the Commissioner heading above.
#cop_config

Configuration Helpers.
#cop_name, #excluded_file?,
#external_dependency_checksum

This method should be overridden when a cop’s behavior depends on state that lives outside of these locations:
#inspect,
#message

Gets called if no message is specified when calling add_offense or add_global_offense Cops are discouraged to override this; instead pass your message directly.
#name

Alias for Base#cop_name.
#offenses,
#on_investigation_end

Called after all on_…​
#on_new_investigation

Called before all on_…​
#on_other_file

Called instead of all on_…​
#parse

There should be very limited reasons for a Cop to do it’s own parsing.
#parser_engine,
#ready

Called between investigations.
#relevant_file?,
#target_gem_version

Returns the version of a gem in the target, by default its locked version (i.e. from Gemfile.lock or gems.locked).
#target_rails_version, #target_ruby_version, #annotate, #apply_correction, #attempt_correction,
#callback_argument

Reserved for Cop::Cop.
#complete_investigation

Called to complete an investigation.
#correct, #covering_disabled_range, #current_corrector,
#current_offense_locations

Reserved for Commissioner:
#current_offenses, #currently_disabled_lines, #custom_severity, #default_severity, #disable_uncorrectable, #enabled_line?,
#enabled_lines?

A multi-line offense is suppressed by a directive on any line of its range, not only its first line, matching the intuition that the directive is attached to the offending code.
#file_name_matches_any?, #find_message, #find_severity, #matches_absolute_include_pattern?, #new_offense, #range_for_original, #range_from_node_or_range,
#reset_investigation

Actually private methods.
#suppression_reason

The -- reason on the directive that suppresses offenses on this range, or nil when the directive carries none.
#use_corrector

::RuboCop::Cop::AutocorrectLogic - Included

#disable_offense, #disable_offense_at_end_of_line, #disable_offense_before_and_after, #disable_offense_with_eol_or_surround_comment, #eol_comment, #eol_comment_would_be_inside_literal?, #heredoc_range, #line_with_eol_comment_too_long?, #max_line_length, #multiline_ranges, #multiline_string?,
#range_by_lines

Expand the given range to include all of any lines it covers.
#range_of_first_line, #string_continuation?, #surrounding_heredoc?, #surrounding_percent_array?

::RuboCop::Cop::IgnoredNode - Included

#ignore_node, #ignored_node?, #part_of_ignored_node?, #ignored_nodes

Constructor Details

This class inherits a constructor from RuboCop::Cop::Base

Instance Method Details

#autocorrect(corrector, node, count_node, replacement) (private)

[ GitHub ] 

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 181



def autocorrect(corrector, node, count_node, replacement)
  if replacement.start_with?('!')
    corrector.insert_before(count_node, '!')
    corrector.replace(count_node.loc.selector, 'empty?')
  else
    corrector.replace(count_node.loc.selector, replacement)
  end

  corrector.remove(removal_range(node))
end


  








  

    #block_given_to_count?(count_node)  ⇒ Boolean  (private)
  



  

    
  





  


  [ GitHub ]


  

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 166



def block_given_to_count?(count_node)
  count_node.parent&.any_block_type? || count_node.last_argument&.block_pass_type?
end


  








  

    #count_predicate(node)  
  

  [ GitHub ]


  

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 127



def_node_matcher :count_predicate, <<~PATTERN
  (send
    {
      (any_block $(call !nil? :count) _ _)
      $(call !nil? :count (block-pass _)?)
    }
    {
      :positive? |
      :> (int 0) |
      :!= (int 0) |
      :zero? |
      :== (int 0) |
      :== (int 1) |
      :> (int 1)
    })
PATTERN


  








  

    #on_send(node)  
  

  [ GitHub ]


  

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 144



def on_send(node)
  return unless (count_node = count_predicate(node))

  replacement = replacement_method(node, count_node)

  return unless replacement_supported?(node, replacement)

  offense_range = count_node.loc.selector.join(node.source_range.end)
  add_offense(offense_range, message: format(MSG, prefer: replacement)) do |corrector|
    autocorrect(corrector, node, count_node, replacement)
  end
end


  








  

    #receiver_of_call?(node)  ⇒ Boolean  (private)
  



  

    
  





  


  [ GitHub ]


  

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 177



def receiver_of_call?(node)
  node.parent&.call_type? && node.parent.receiver == node && node.parent.loc.dot
end


  








  

    #removal_range(node)   (private)
  

  [ GitHub ]


  

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 192



def removal_range(node)
  range = (node.loc.dot || node.loc.selector).join(node.source_range.end)

  range_with_surrounding_space(range, side: :left)
end


  








  

    #replacement_method(node, count_node)   (private)
  

  [ GitHub ]


  

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 159



def replacement_method(node, count_node)
  key = [node.method_name, node.first_argument&.value]
  return REPLACEMENTS.fetch(key) if block_given_to_count?(count_node)

  NO_BLOCK_REPLACEMENTS[key]
end


  








  

    #replacement_supported?(node, replacement)  ⇒ Boolean  (private)
  



  

    
  





  


  [ GitHub ]


  

  


# File 'lib/rubocop/cop/style/collection_querying.rb', line 170



def replacement_supported?(node, replacement)
  return false unless replacement
  return false if replacement == 'many?' && !active_support_extensions_enabled?

  !(replacement.start_with?('!') && receiver_of_call?(node))
end