We are creating a graph database, and we are interested in how to model the structure of the graph. The idea is to have users who can relate to several things (friends, blog posts, etc.). Our doubt is that it is better to have intermediate nodes like this:
User1 -+-> Friends -+-> User2
| |
| +-> User3
|
+-> Posts -+-> Post1
|
+-> Post2
Or, if you directly connect the user to your elements, for example:
User1 -+-> User2
|
+-> User3
|
+-> Post1
|
+-> Post2
We see the benefits and challenges of both styles. It would be great to hear the opinions of people with extensive experience working with social graphs.
source
share