What experiments to you do on very large graphs?

[Amirouche Boubekki]

Hello,

I am working on a graph database in Python with several backends

including backends that persist the graph to the disk (based on leveldb),

that is, it can handle bigger than RAM graphs.

I am wondering what people use right now for working

with very large graphs? neo4j? tinkerpop? graph-tool?

If you can not say the exact algorithms you use.

Can you at least share the domain or subdomain.

What would be the requirements for you to use such a library?

Thanks in advance!

[Russell Jurney]

For very large graphs, GraphFrames on PySpark is indispensable. Many people use the connected components, but my favorite part is the network motif search.

Docs: https://graphframes.github.io/graphframes/docs/_site/index.html
Code: https://github.com/graphframes/graphframes

Hope this helps,

Russ