|
As far as I understand,
you will see differences on a huge amount of nodes in relation of update/reads operations.
For example if I you have to remove/add elements, our current approach should be faster because we just add/remove a node.
using a linked list you also have to update the relationships.
If you need to traverse the list to read a particular element than the linked list approach is much faster.
Basically, Neo4j is very fast when it comes to traverse nodes and when dealing with it you choose your representation based on the query you need to run.
As for the null values,
We can add an annotation that says
for those cases where the user does not care about the null values,
Actually a better solution could be
.
It would make sense since Neo4j does not store null values and we might want a similar annotation anyway.
|
