rails: Giant queries

Giant_queries.md

Giant queries loading everything into memory

You need to fix some data so you’ll just iterate through it all and fix it, right?

User.has_purchased(true).each do |customer|
  customer.grant_role(:customer)
end

You have an ecommerce site with a million customers. Let’s say each User object takes 500 bytes. This code will take 500MB of memory at runtime! Better:

User.has_purchased(true).find_each do |customer|
  customer.grant_role(:customer)
end

find_each uses find_in_batches to pull in 1000 records at a time, dramatically lowering the runtime memory requirements.

Bonus points: use update_all or raw SQL to perform the mass update. SQL takes time to learn well but the benefits are even more tremendous: you’ll see a 100x improvement in the performance.