Difference in implementation between clojure.core conj and clojure.core.into

Question

What is the difference in terms of efficiency and under-the-hood details

between (apply conj [0][1]) and (into [0][1])?

Using the time function it seems that apply conj is faster, but is there a price? and what is it?

Answer 1

The into function is generally preferable to apply conj in Clojure due to the following reasons:

• Performance: into is usually faster than apply conj. The into function is implemented in a way that allows for optimized, efficient concatenation of collections. On the other hand, apply conj involves using conj repeatedly with variadic arguments, which can be less performant.

• Simplicity & Readability: into provides a simpler and more concise way to concatenate collections. It takes care of the concatenation internally without requiring explicit iteration or multiple function calls. Using into makes the code more readable and expressive. It clearly communicates the intention of concatenating or merging collections, improving the overall code clarity.

• Flexibility: into supports merging of collections, not just concatenation. It can merge collections of different types or with conflicting keys, providing more flexibility in combining data structures.

That being said, there may be specific cases where apply conj is more suitable, such as when you need to dynamically pass in collection elements as variadic arguments. However, for most general cases, the into function is the preferred way to concatenate or merge collections in Clojure.