SemyonSinchenko commented on code in PR #58302: URL: https://github.com/apache/spark/pull/58302#discussion_r4284525425
########## graphframes/src/main/scala/org/apache/spark/graphframes/propertygraph/PropertyGraphFrame.scala: ########## Review Comment: What do you think about moving out the current LPG (labeled property graph) API at all from the scope of this PR? My proposal is let's migrate GraphFrames without LPG for now, then I will prepare a better way of this API. The main reason of its existence was to unblock the ISO GQL / `MATCH` API. Without real query capabilities the LPG API does not make a lot of sense. And due to the more strict policy about breaking changes in Spark compared to GraphFrames, I would prefer to rethink this API from scratch. The perfect scenario would be to cache the graphs in the similar way like ML works over Spark Connect. And the design needs to be reviewed more carefully. Also if the Spark project does not want to accept the ISO GQL `MATCH` API at all, the LPG API is redundant as well. What do you think about moving this part out of the scope of the initial migration? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
