On Fri, Nov 4, 2011 at 10:17 PM, maxdemarzi <maxdema...@gmail.com> wrote:
> > Tero Paananen wrote: > > > > Additionally I don't find adding a join keyword to a query language that > > queries a data store that has no joins better in any shape or form. > > > > That is one way of looking at it, another way of looking at it is that all > the tables are already joined. > Not sure how we can say a graph database "has no joins". > > Now, we don't have to throw out anything. If this is pie in the sky talk > anyway, we can have it both ways. > (1) lucy-[:ACTS_IN]->movie > (2) lucy-[:ACTS_IN]-movie > (3) lucy<-[:ACTS_IN]-movie > > (4) lucy out(:acts_in) movie > (5) lucy both(:acts_in) movie > (6) lucy in(:acts_in) movie > > What I like about spelling out the relationship direction is: > A. You avoid typo errors. Was it -- or ->, look how close 1 and 2 are. > B. You know right away which direction you are going. Compare 1 and 2, vs > 4 > and 5. At the 6th character in both 4 and 5, you know which direction it > is. You have to get to the 17th character in 1 and 2. > C. We read left to right so: > -- 1 reads: lucy via relationship acts_in outgoing to movie, > -- 2 reads: lucy via relationship acts_in both to movie > -- 3 reads, lucy incoming via relationship acts_in to movie > That is inconsistent. > Try: > -- 4 reads, lucy outgoing via relationshp acts_in to movie > -- 5 reads, lucy both via relationshp acts_in to movie > -- 6 reads, lucy incoming via relationshp acts_in to movie > D. try (a)-->(b)-->(c)-->(d)--(e)-->(f)-->(g) #did you catch the both > relationship between d and e? or did your eyes scan over it? If you're > debugging someone else's code, was it a typo or did they intend it that > way? > If you spell it out there is no ambiguity. > > Anyway, if the --> syntax is polarizing (as it seems to be) then *give > people the choice* to use whichever they prefer. Love the idea. Why didn't I think of that? All of Cypher is made out of bits and pieces that I've stolen from others anyway - I think I want to steal this one too. Being able to have these discussions with our users is a luxury that I'm very happy about - thank you! Andrés _______________________________________________ Neo4j mailing list User@lists.neo4j.org https://lists.neo4j.org/mailman/listinfo/user
