In the course of working on a tree-like data structure, I have designed a struct for each node in the tree. Since there are only 5 fields in each node, and all nodes have the same 5 fields, using a struct made sense to me.

However, one of the %Node{} struct fields contains the children of the node, which is a list of %Node{} structs, each of which can have its own children, and so on.

Given the challenges of accessing nested structs, I am beginning to question my original decision to use a struct instead of a map. I see that there are a number of posts already on how to access nested structs, but I’m still left asking: What does a struct really offer me here? Basically, it offers me the ability to limit the keys for each node. But is that a worthwhile tradeoff, given that the number of nested levels could be very high? Is it really a problem to just use a map?

The nature of the code is such that reaching down many levels at once is unlikely to be needed. For the most part, the tree will be walked one level at a time, but it will frequently be necessary to match on or modify fields in the node’s children (1 level down). It seems that the risk of unintended fields on the struct is perhaps not worth the complexity of using a struct in the first place.

Thoughts?