R Programming By Example

Preallocating memory to avoid duplication

Memory preallocation is an important technique we covered implicitly when we used the lapply() function, since it does preallocation for us. However, a more explicit explanation can be useful. As we have already seen, dynamically growing objects in R is not great for performance. Instead, you should define an object with the full size you will need and then perform updates on its elements instead of recreating them. To accomplish this, you may use something like double(10) to define an vector of doubles that will contain 10 elements at most. Whenever you define an object's size before you start using it, will help you avoid recreating new objects each time its size is increased and will save you a ...