Sure… don’t eager load. Get your first result set… then use collect to grab the ids.
ids_to_find = @results.collect{|r| r.id}
Then you can use that array in another finder.
either like this:
more_results = Result.find(ids_to_find)
or
more_results = Result.find(:all, include =>[:other_stuff], :conditions=>[“results.id in (?)”, ids_to_find]
Would that work?
I have created a plugin (my first) which monkey patches the functionality submitted in http://dev.rubyonrails.org/ticket/7147 into ActiveRecord, with several modifications.
It has been modified to use a new naming convention for :include relationships.
Instead of using t[table_index]_r[column_index] I am now using a more human readable convention of the form table_name[index]__column_name.
Also, One of our requirements was the ability to extract an ActiveRecord object hierarchy from an arbitrary (albeit following the naming convention) result set. This is now possible through the extract_objects(rows) method. We had several find methods which had to be converted to PL/SQL functions for performance reasons. We can now use connection.select_all to grab a result set, and run it through extract objects to get a friendly active record collection (works with nested includes as well)
I have not written comprehensives tests for it yet (i know, bad!) But the source could be useful so I have decided to release it on RubyForge at http://rubyforge.org/frs/?group_id=4193
It has been working well thus far in my application, YMMV. Please give me feedback if you find any problems or have improvements.
Thanks,
Matt Kull