When I began learning rails I found it frustrating that array and active record relation objects could not be queried in the same fashion as database tables. Before too much confusion let me explain.
When I want something from a database table I simply write something like: Product.where(:color=>‘black’) NOW, if I put a .to_sql after that statement I see something like “SELECT "styles".* FROM "styles" WHERE "styles"."color" = ‘black’”
So clearly, rails is magically converting that friendly statement into something more meaningful to the backend. That’s fantastic and makes my life wonderful.
So here’s my problem… after running the query I am returned an active record relation, which let’s say is called @products. Now, let’s say I want to further break that list of products into products that are big and small. This would require one of two things… either I must hit the database twice being more selective, or I must iterate over @products using a block and such ruby magic as collect/map/select/etc…
To me neither of the above is a great option, and I must admit while learning ruby on rails this was one of the biggest stumbling blocks for me, and one that has produced a massive amount of stackoverflow headaches for many others.
My suggestion to rails developers is to make rails more accessible to beginners by allowing programmers to apply the same simplicity of writing an active record query to querying an active record relation object or array of model objects.
For example, I want to now take @products and say:
@products.where(:size=>“small”) and have this converted automagically by rails to the appropriate statement @products.select{|product| product.size == ‘small’}
I suppose this could work if rails were to query the object type first before performing the query. Is the object type an array or active record relation, and is it in the expected format (collection of appropriate objects)? If it is, then treat it as a virtual table and instead of performing an sql query, hit it with a select block, etc.
I know understanding the underlying ruby code is good, and at times would be preferable to using active record style commands… but for new users getting off the ground, I think having both options would be beneficial. Also, I feel it would be more consistency to the framework generally.
I have also posted this at reddit http://www.reddit.com/r/rails/comments/12kyho/suggestion_for_rails_4_what_are_pros_and_cons/