Hello!

I have already posted a similar thing on stackoverflow and i mean in no way 
to cross-post the actual problem (which is all made up btw)  but rather 
understand why a graph algorithm couldnt be useful when solving certain 
problems
http://stackoverflow.com/questions/22646305/is-there-anything-like-a-do-while-match-pattern-that-satisfy-an-aggregated-val

So, i posted that before i even knew there was something called knapsack 
problem:
http://rosettacode.org/wiki/Knapsack_problem/0-1

where a problem is forumlated as
*Which items does the tourist carry in his knapsack so that their total 
weight does not exceed 400 dag [4 kg], and their total value is maximised?*
I think this question combined with a graph pattern is useful becuase the 
traversal is actually taking place so the cost could be calculated.

And in many ways this is much similar to what i am trying to understand how 
graph query can be used or am i wrong on this ? Should it be on application 
level only?

What i eventually would like to do is to test the reduce/accumulator 
function or limit not only on rows but on actual property values if that 
make sense.

Assume we have a graph with people and page nodes. The page nodes have 
stats on them. This query accumulates the stats for the pages a person has 
visited

MATCH (n:People)-[:VISITS]-sites WITH reduce(acc=0, x IN collect(sites.
dailyhits)| acc + x) AS totalhits RETURN totalhits;


I thinks an expression like this could be useful to limit the pattern:

MATCH (n:People)-[:VISITS]-sites WITH reduce(acc=0, x IN collect(sites.
dailyhits)| acc + x) AS totalhits WHERE totalhits<50000 RETURN totalhits;

I understand why it doesnt work, but the idea of evaluate an expression for 
each graph traverse or compare paths seems to be a very useful way of 
getting a more appropriate resultsets for some problems.

Any inputs or general ideas ?

-- 
You received this message because you are subscribed to the Google Groups 
"Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to