Fetching contributors…
Cannot retrieve contributors at this time
228 lines (166 sloc) 9.38 KB


Status Gem Version

A simple, extensible Ruby client for Apache Solr.


The code docs


gem install rsolr


require 'rsolr'

# Direct connection
solr = RSolr.connect :url => ''

# Connecting over a proxy server
solr = RSolr.connect :url => '', :proxy=>''

# Using an alternate Faraday adapter
solr = RSolr.connect :url => '', :adapter => :em_http

# Using a custom Faraday connection
conn = do |faraday|
  faraday.response :logger                  # log requests to STDOUT
  faraday.adapter  Faraday.default_adapter  # make requests with Net::HTTP
solr = RSolr.connect conn, :url => ''

# send a request to /select
response = solr.get 'select', :params => {:q => '*:*'}

# send a request to /catalog
response = solr.get 'catalog', :params => {:q => '*:*'}

When the Solr :wt is :ruby, then the response will be a Hash. This Hash is the same object returned by Solr, but evaluated as Ruby. If the :wt is not :ruby, then the response will be a String.

The response also exposes 2 attribute readers (for any :wt value), :request and :response. Both are Hash objects with symbolized keys.

The :request attribute contains the original request context. You can use this for debugging or logging. Some of the keys this object contains are :uri, :query, :method etc..

The :response attribute contains the original response. This object contains the :status, :body and :headers keys.

Request formats

By default, RSolr uses the Solr JSON command format for all requests.

RSolr.connect :url => '', update_format: :json # the default
# or
RSolr.connect :url => '', update_format: :xml


The read and connect timeout settings can be set when creating a new instance of RSolr:

solr = RSolr.connect(:read_timeout => 120, :open_timeout => 120)

Retry 503s

A 503 is usually a temporary error which RSolr may retry if requested. You may specify the number of retry attempts with the :retry_503 option.

Only requests which specify a Retry-After header will be retried, after waiting the indicated retry interval, otherwise RSolr will treat the request as a 500. You may specify a maximum Retry-After interval to wait with the :retry_after_limit option (default: one second).

solr = RSolr.connect(:retry_503 => 1, :retry_after_limit => 1)

For additional control, consider using a custom Faraday connection (see above) using its `retry` middleware.


Use the #get / #post method to send search requests to the /select handler:

response = solr.get 'select', :params => {
response["response"]["docs"].each{|doc| puts doc["id"] }

The :params sent into the method are sent to Solr as-is, which is to say they are converted to Solr url style, but no special mapping is used. When an array is used, multiple parameters *with the same name* are generated for the Solr query. Example:

solr.get 'select', :params => {:q=>'roses', :fq=>['red', 'violet']}

The above statement generates this Solr query:



To paginate through a set of Solr documents, use the paginate method:

solr.paginate 1, 10, "select", :params => {:q => "test"}

The first argument is the current page, the second is how many documents to return for each page. In other words, “page” is the “start” Solr param and “per-page” is the “rows” Solr param.

The paginate method returns WillPaginate ready “docs” objects, so for example in a Rails application, paginating is as simple as:

<%= will_paginate @solr_response["response"]["docs"] %>

Method Missing

The +RSolr::Client+ class also uses method_missing for setting the request handler/path:

solr.paintings :params => {:q=>'roses', :fq=>['red', 'violet']}

This is sent to Solr as:


This works with pagination as well:

solr.paginate_paintings 1, 10, {:q=>'roses', :fq=>['red', 'violet']}

Using POST for Search Queries

There may be cases where the query string is too long for a GET request. RSolr solves this issue by converting hash objects into form-encoded strings:

response = :data => {:q => "*:*"}

The :data hash is serialized as a form-encoded query string, and the correct content-type headers are sent along to Solr.

Sending HEAD Requests

There may be cases where you'd like to send a HEAD request to Solr:

solr.head("admin/ping").response[:status] == 200

Sending HTTP Headers

Solr responds to the request headers listed here: To send header information to Solr using RSolr, just use the :headers option:

response = solr.head "admin/ping", :headers => {"Cache-Control" => "If-None-Match"}

Building a Request

+RSolr::Client+ provides a method for building a request context, which can be useful for debugging or logging etc.:

request_context = solr.build_request "select", :data => {:q => "*:*"}, :method => :post, :headers => {}

To build a paginated request use build_paginated_request:

request_context = solr.build_paginated_request 1, 10, "select", ...

Updating Solr

Updating is done using native Ruby objects. Hashes are used for single documents and arrays are used for a collection of documents (hashes). These objects get turned into simple XML “messages”. Raw XML strings can also be used.

Single document via #add

solr.add :id=>1, :price=>1.00

Multiple documents via #add

documents = [{:id=>1, :price=>1.00}, {:id=>2, :price=>10.50}]
solr.add documents

The optional :add_attributes hash can also be used to set Solr “add” document attributes:

solr.add documents, :add_attributes => {:commitWithin => 10}

Raw commands via #update

solr.update data: '<commit/>', headers: { 'Content-Type' => 'text/xml' }
solr.update data: { optimize: true }.to_json, headers: { 'Content-Type' => 'application/json' }

When adding, you can also supply “add” xml element attributes and/or a block for manipulating other “add” related elements (docs and fields) by calling the xml method directly:

doc = {:id=>1, :price=>1.00}
add_attributes = {:allowDups=>false, :commitWithin=>10}
add_xml = solr.xml.add(doc, add_attributes) do |doc|
  # boost each document
  doc.attrs[:boost] = 1.5
  # boost the price field:
  doc.field_by_name(:price).attrs[:boost] = 2.0

Now the “add_xml” object can be sent to Solr like:

solr.update :data => add_xml


Delete by id

solr.delete_by_id 1

or an array of ids

solr.delete_by_id [1, 2, 3, 4]

Delete by query:

solr.delete_by_query 'price:1.00'

Delete by array of queries

solr.delete_by_query ['price:1.00', 'price:10.00']

Commit / Optimize

solr.commit, :commit_attributes => {}
solr.optimize, :optimize_attributes => {}

Response Formats

The default response format is Ruby. When the :wt param is set to :ruby, the response is eval'd resulting in a Hash. You can get a raw response by setting the :wt to +“ruby”+ - notice, the string – not a symbol. RSolr will eval the Ruby string ONLY if the :wt value is :ruby. All other response formats are available as expected, +:wt=>'xml'+ etc..

Evaluated Ruby:

solr.get 'select', :params => {:wt => :ruby} # notice :ruby is a Symbol

Raw Ruby:

solr.get 'select', :params => {:wt => 'ruby'} # notice 'ruby' is a String


solr.get 'select', :params => {:wt => :xml}

JSON (default):

solr.get 'select', :params => {:wt => :json}

Related Resources & Projects

Note on Patches/Pull Requests

  • Fork the project.

  • Make your feature addition or bug fix.

  • Add tests for it. This is important so I don't break it in a future version unintentionally.

  • Commit, do not mess with rakefile, version, or history (if you want to have your own version, that is fine but bump version in a commit by itself I can ignore when I pull)

  • Send me a pull request. Bonus points for topic branches.


  • Nathan Witmer

  • Magnus Bergmark

  • shima

  • Randy Souza

  • Mat Brown

  • Jeremy Hinegardner

  • Denis Goeury

  • shairon toledo

  • Rob Di Marco

  • Peter Kieltyka

  • Mike Perham

  • Lucas Souza

  • Dmitry Lihachev

  • Antoine Latter

  • Naomi Dushay


Matt Mitchell <>


Copyright © 2008-2010 Matt Mitchell. See LICENSE for details.