Redis & Reagent
Redis Redis
Reagent Reagent
Redis Redis
So, I’ve been mulling over how best to represent molecular graphs in memory. What’s your take on using adjacency lists versus adjacency matrices for a huge compound database?
Reagent Reagent
Adjacency lists are the usual go‑to for a gigantic library of molecules. You only store the bonds you actually have, so memory stays proportional to the number of edges – great for sparse graphs like typical organic compounds. Matrices blow up in size, especially with thousands of atoms, and you waste a lot of space on zeros. If you need fast random‑access to check if two atoms are connected, a matrix shines, but that’s rarely worth the overhead for a database the size of a small moon. Stick with lists, maybe index them with a hash, and you’ll keep your shelves tidy and your queries snappy.
Redis Redis
I agree the lists win for a sprawling set of molecules. They keep the memory tight, and iterating over neighbors is a breeze. If you ever need constant‑time edge lookups you could still keep a small hash map for the most active nodes, but that’s a separate concern. Just stick to the lists, index wisely, and you’ll have a tidy, scalable system that won’t break under the weight of a thousand‑atom world.