Man’s Search for Meaning

Counter caching single columns is an excellent feature in Rails [1]. However, the need to counter cache values from multiple columns occurs frequently in complex real world applications. Let’s explore a non database solution.

Here’s a toy example :

To keep track of all the products in a store we have the products_count column which rails automatically updates because we set :counter_cache => true

Counter caching quickly becomes inadequate if we have to store the total price of all products in a store.  Of course, we can easily implement our own.

Now, suppose we have to keep track of the total price and total count of products in every category in every store. It’s a nightmare to develop and maintain if each value is kept in the database.

Enter caching. I recommend Amazon’s ElastiCache [2] or any implementation of memcached. Though, Rails can cleanly accomodate most caching solutions [3].

But first, make sure you have indexes.

Then, convert counter caches into queries and place them in cache blocks.

And, of course delete cache entries in a product observer.

That’s it. By using cache storage instead of database storage for complex counter caching the code becomes simpler, maintainable and scalable. 

[1] Counter Cache Column (Railscasts)

[2] Amazon ElastiCache

[3] Caching in Rails (Railscasts)