Class: Parse::Constraint::DoesNotEqualLinkedPointerConstraint

Inherits:

Constraint

• Object
• Constraint
• Parse::Constraint::DoesNotEqualLinkedPointerConstraint

Defined in:

lib/parse/query/constraints.rb

Overview

Constraint for comparing pointer fields where they do NOT equal through linked objects. Uses MongoDB's $lookup to join collections and $expr with $ne to compare fields.

Usage:

Document.where(:project.does_not_equal_linked_pointer => { through: :post, field: :project })

This generates a MongoDB aggregation pipeline that:

1. Uses $lookup to join the linked collection
2. Uses $match with $expr and $ne to find records where fields do NOT match

Examples:

Find assets where the project does not equal the post's project

Document.where(:project.does_not_equal_linked_pointer => {
  through: :post,
  field: :project
})

Instance Method Summary

• #build ⇒ Hash

  Builds the MongoDB aggregation pipeline for the does-not-equal-linked-pointer constraint.

Instance Method Details

#build ⇒ Hash

Builds the MongoDB aggregation pipeline for the does-not-equal-linked-pointer constraint

Returns:

• (Hash) —

  Hash containing the aggregation pipeline

Raises:

• (ArgumentError) —

  if required parameters are missing or invalid

# File 'lib/parse/query/constraints.rb', line 3162

def build
  # Validate that value is a hash with required keys
  unless @value.is_a?(Hash) && @value[:through] && @value[:field]
    raise ArgumentError, "DoesNotEqualLinkedPointerConstraint requires a hash with :through and :field keys"
  end

  through_field = @value[:through]
  target_field = @value[:field]

  # Convert field names to Parse format (snake_case to camelCase) with _p_ prefix for pointers
  local_field_name = format_field_name(@operation.operand, is_pointer: true)
  through_field_name = format_field_name(through_field, is_pointer: true)
  target_field_name = format_field_name(target_field, is_pointer: true)

  # Determine the collection name for the lookup (Rails pluralization)
  through_class_name = through_field.to_s.classify
  lookup_collection = through_class_name

  # Generate unique alias name for the joined data (use clean name without _p_ prefix)
  lookup_alias = "#{through_field.to_s.camelize(:lower)}_data"

  # Build the MongoDB aggregation pipeline
  pipeline = []

  # Parse stores pointers as "ClassName$objectId" strings
  # We need to extract just the objectId part after the $
  # Stage 1: Add field with extracted objectId
  add_fields_stage = {
    "$addFields" => {
      "#{through_field_name}_id" => {
        "$substr" => [
          "$#{through_field_name}",
          lookup_collection.length + 1,  # Skip "ClassName$"
          -1,  # Rest of string
        ],
      },
    },
  }
  pipeline << add_fields_stage

  # Stage 2: $lookup to join the linked collection
  lookup_stage = {
    "$lookup" => {
      "from" => lookup_collection,
      "localField" => through_field_name,
      "foreignField" => "_id",
      "as" => lookup_alias,
    },
  }
  pipeline << lookup_stage

  # Stage 2: $match with $expr to compare the fields using $ne (not equal)
  match_stage = {
    "$match" => {
      "$expr" => {
        "$ne" => [
          { "$arrayElemAt" => ["$#{lookup_alias}.#{target_field_name}", 0] },
          "$#{local_field_name}",
        ],
      },
    },
  }
  pipeline << match_stage

  # Return a special marker that indicates this needs aggregation pipeline processing
  { "__aggregation_pipeline" => pipeline }
end