CODEWITHSUNDEEP
ruby-on-rails-3.2elasticsearchthinking-sphinxtire
Nazar Hussain
Nazar Hussain

Reputation: 5162

Alternative syntax for ElasticSearch vs ThinkingSphinx

Please have a look on below code, Its the normal indexing statement for thinking sphinx

indexes owned_tags.name, :as => :owned_tag_names
has owned_tags.id, :as => :owned_tag_ids, :facet => true

Can any one guide what would be the syntax to have same index in ElasticSearch ?

I tried out. 

tire.mapping do
  indexes :owned_tag_names, type: 'array', analyzer: :ngram_analyzer, index_name: 'tag'
  indexes :owned_tag_ids,   type: 'array'
end

def to_indexed_json
{
    owned_tag_names: owned_tags.collect(&:name),
    owned_tag_ids: owned_tags.collect(&:id)
}.to_json
end

and it is giving me the error: 

400 : {"error":"MapperParsingException[mapping [user]]; nested: MapperParsingException[No handler for type [array] declared on field [owned_tag_ids]]; ","status":400}

Upvotes: 0

Views: 994

Answers (1)

karmi
karmi

Reputation: 14419

Use the string type for definining mapping for arrays, as the documentation suggests:

tire.mapping do
  indexes :owned_tag_names, type: 'string', analyzer: :ngram_analyzer, index_name: 'tag'
  indexes :owned_tag_ids,   type: 'string'
end

Note, that you could define the JSON serialization in the mapping block with the as option:

tire.mapping do
  indexes :owned_tag_names,
          type: 'array',
          analyzer: :ngram_analyzer,
          index_name: 'tag',
          as: proc { owned_tags.collect(&:name) }

  indexes :owned_tag_ids,
          type: 'array',
          as: proc { owned_tags.collect(&:id) }
end

Upvotes: 1

Related Questions