If you want to help hack on this, I've written a Postgres backend for the graph database Cayley that's fairly fast and about to merge (for use on a prod feature): https://github.com/google/cayley/pull/289
For traversals, it tries its level best to let Postgres create the appropriate JOINs, which Postgres does well... most of the time anyway. In doing this, I learned really fast how many weird semantic corners of bog-standard SQL are, and the optimization choices Postgres makes based on the way you formulate the query.
Ping me on that PR! And/or join #cayley on Freenode
Comments
If you want to help hack on this, I've written a Postgres backend for the graph database Cayley that's fairly fast and about to merge (for use on a prod feature): https://github.com/google/cayley/pull/289
For traversals, it tries its level best to let Postgres create the appropriate JOINs, which Postgres does well... most of the time anyway. In doing this, I learned really fast how many weird semantic corners of bog-standard SQL are, and the optimization choices Postgres makes based on the way you formulate the query.
Ping me on that PR! And/or join #cayley on Freenode
Thanks for sharing this! I would love to help with this - once I have gotten up to speed on various things by building a graph on Postgres myself.