Marko,Thanks for the additional insights, a WeightedPath in the neo4j world is a path that contains weights between nodes, in the realm of calculating the shortest path using for example the Dijkstra Algorithm the path weights (in this case the distances) are all summed up and the path with the minimum total weight is the one returned. Here's an example using the weightedpath in the neo4j world (http://components.neo4j.org/neo4j-graph-algo/). In my graph all paths have weights corresponding to length, that being said how do I then convert the data returned by Gremlin to a datastructure that represents a list of WeightedPath objects. I need this datastructure to be returned through json back to the client. Thanks for the other links, I'll look through them soon and look forward to your blog. Regards
> From: [email protected] > Date: Thu, 17 Feb 2011 15:53:48 -0700 > To: [email protected] > Subject: Re: [Neo4j] Extracting paths from a subgraph > > Hi, > > > 1) I tried a more complicated version of the query below and it worked > > fine, my question here is does this return a list of WeightedPath objects > > in the neo4j world, if not how do I convert the list of paths from the > > gremlin world to the neo4j world, if I have to convert is this trivial and > > are there examples? > > I don't know what WeightedPaths are in Neo4j. Is that the sum of the weights > on edges of a path? Can you give me some more information? > > > 2) I watched your screencast, it was awfully fast :) and enlightening, have > > you uploaded some more involved code examples (in java and potentially > > groovy) of graph searching like the query below, that would be priceless. > > There are lots of places to find snippets: > https://github.com/tinkerpop/gremlin/wiki (Gremlin documentation has > lots of examples throughout) > http://markorodriguez.com/category/blog/ (my blog) > https://github.com/tinkerpop/gremlin/wiki/Gremlin-Cookbook (user > provided recipes) > > If you have particular examples you would like to see in the Cookbook, I can > always add them. The play graph we use in the Gremlin documentation is > defined here > https://github.com/tinkerpop/gremlin/wiki/Defining-a-More-Complex-Property-Graph > .. If you tell me a traversal problem for that domain model, I will happily > solve it and provide it as a recipe. > > > 3) Are there javadoc equivalents to learn and use Gremlin, I didnt see > > anything and figured that it only has a few reserved words needed to query > > and extract data from graphs > > Here are the "Cheat Sheets": > https://github.com/tinkerpop/gremlin/wiki/Gremlin-Steps > https://github.com/tinkerpop/gremlin/wiki/Gremlin-Methods > > > 4) Have you done any load testing running these kinds of DSL based > > expression queries on large graphs and if so what are the results of some > > of these tests? > > Gremlin is a lazy graph traversal framework that works over various GraphDBs > (Neo4j, OrientDB, various Sail, soon InfiniteGraph, and less soon DEX). The > largest Neo4j graph that I personally have used Gremlin over was 10 million > edges (?? vertices). An empirical proof that Blueprints > (http://blueprints.tinkerpop.com) is nearly identical in speed to raw Neo4j > is here (Blueprints is what Gremlin talks too): > http://groups.google.com/group/gremlin-users/msg/c94dfef8352f68d3 > > I'm currently working on new blog post that shows how such performance > characteristics for the GraphDBs currently supported by Blueprints. That > should be out next week with a new round of releases for the TinkerPop stack. > > > Thanks again for all your help.Regards > > No problem. Enjoy Gremlin!, > Marko. > > http://markorodriguez.com > _______________________________________________ > Neo4j mailing list > [email protected] > https://lists.neo4j.org/mailman/listinfo/user _______________________________________________ Neo4j mailing list [email protected] https://lists.neo4j.org/mailman/listinfo/user
