The ELC Community Blog

A knowledge exchange on Ruby on Rails and Agile Development

Advanced db_free_solr

by jmoline on March 31, 2008

If you've implemented the db_free_solr plugin in your application, it's probably safe to assume that performance is pretty important to you. For those of you that are looking to squeeze just a little bit more efficiency out of your Solr implementation, here's another feature you might want to know about.

The out-of-box functionality that db_free_solr provides, much the same as acts_as_solr, is designed to make things as easy as possible for you. To that end, the default behavior is to index and store every field you call out in your models. The truth, though, is that you won't need every field you index to be returned with the results. Likewise, some of the fields you want to have returned will, no doubt, not be necessary to index for searching purposes.

So, how do we go about avoiding unnecessary indexing and storing?

The answer is actually simpler than you might think. Acts_as_solr allows you to specify the datatypes of the fields you tell it to index (if not specified, the index assumes a datatype of text). Like so:

acts_as_solr :fields => [{:field1 => :string}, {:field2 => :integer}, {:field3 => :date}]

So, what I've done, is to add a few datatypes to correspond with the newly available indexing options. As I said above, the default is to both index and store, so what I've added are datatypes to indicate that a field should be marked for storage only ("_so") or indexing only ("_io"). These can be used like so:

acts_as_solr :fields => [{:field1 => :string_io}, {:field2 => :integer_so}, {:field3 => :date}]

It's also worth noting that you can specify datatypes for each and every field, or only a select few. In other words, this will work just fine:

acts_as_solr :fields => [:field1, {:field2 => :integer_so}, :field3]

Here's a complete listing of the supported datatypes:

Symbol	Definition
:text	text, indexed and stored
:text_so	text, stored only
:text_io	text, indexed only
:string	string, indexed and stored
:string_so	string, stored only
:string_io	string, indexed only
:date	date, indexed and stored
:date_so	date, stored only
:date_io	date, indexed only
:integer	integer, indexed and stored
:integer_so	integer, stored only
:integer_io	integer, indexed only
:float	float, indexed and stored
:float_so	float, stored only
:float_io	float, indexed only
:boolean	boolean, indexed and stored
:boolean_so	boolean, stored only
:boolean_io	boolean, indexed only

That's all for now. Happy Programming!