Hello, 

I have a similar question to:
https://groups.google.com/d/msg/orient-database/UNPnYPgpAzU/RD1vZrC9AXsJ

but since 4 years are passed, maybe there is something new to consider.

Here using:
Orientdb 2.1.0

Model: A weighted graph, undirected.
Scope: 
a. Query and paginate neighbors of vertexes
b. Query allshortests paths between two (or more vertexes)
c. Query against full-text index and get first neighbors for each vertex + 
paginate

My questions are about a best way to import my model into the db:
Will double insertion of edges (A>B, B>A) be  an issue to traverse the db ?
Is there any native API *available in python* with pyorient implementing 
*allshortestpaths*? (I am coming from neo4j and would like to try same 
queries: particularly, traversal of a weighted graph in combination with 
lucene indexes and sorting.
Will the* order-in-import *of nodes / edges affect my model and traversal?

*As example,* say I have csv tables with redundant edges (A>B, B>A); if 
double edges are removed, a table list like will be produced.
A (no edge)
B (no edge)
C, (> A, > B)

That means, given a dataset of millions of records, you will have huge 
clusters (in this example, C), pointing out to many many leaves, so the 
topology of the graph-db would result very different from my actual graph 
(smaller hubs, small-world).

I would like to ensure that:
1) order in import won't affect traversal and shortest paths
2) how to query the graph with undirected links
either:
3) if it is best to insert double links anyway:
(that is, should I *CREATE EDGE rel FROM #11 
<https://github.com/orientechnologies/orientdb/issues/11>:1 TO #11 
<https://github.com/orientechnologies/orientdb/issues/11>:2* and also *CREATE 
EDGE rel FROM #11 
<https://github.com/orientechnologies/orientdb/issues/11>:2 TO#11 
<https://github.com/orientechnologies/orientdb/issues/11>:1* ?)




-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" 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