Version controlling the graph


(Jeff Hull) #1

Hi All,

I have a use case where I would like to revert the graph back to a point in time…basically time traveling the entire history of the graph.

Is this something that would be totally ridiculous to think about doing, or is there some way to do it that wouldn’t take up a huge amount of space and/or processing power? Maybe this would play with the new transactions in badger, so that each transaction is a point in time that can be reverted to and queried as if it were the latest graph?

I’m thinking like git…except with no need for multiple branches…just one master branch all the way back to the inception of the graph. Presumably this would hose the indexes which are based on the latest data, but maybe there is some way to get around this by viewing history for only some small region of the graph.


Version controlling a predicate
(Manish R Jain) #2

This might be possible – we’re exploring the idea of having MVCC for Dgraph (this is independent of Badger). No word on when that would be though.


(Lazhar Ichir) #3

there was a whitepaper that i cannot find again which added a set of guidelines to RDF for handling versioning properly with quads. will try research it again.

some other papers worth mentioning:


(system) #4

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.