Edge.rank order

Users Issues & Bugs
#1

The edge with a higher rank value comes first.
Data Model - Nebula Graph Database Manual

Statement:

GO FROM "11111111111111111111" OVER follow yield follow._dst, follow._rank | LIMIT 5

Ordered asc:

+------------------------+--------------+
| follow._dst            | follow._rank |
+------------------------+--------------+
| "22222222222222222222" | 1236450693   |
+------------------------+--------------+
| "33333333333333333333" | 1264510802   |
+------------------------+--------------+
| "44444444444444444444" | 1282999563   |
+------------------------+--------------+
| "55555555555555555555" | 1318782385   |
+------------------------+--------------+
| "66666666666666666666" | 1369086354   |
+------------------------+--------------+
#2

@doublemax Thanks for bringing this issue! What version of Nebula Graph are you using?

#3

@doublemax Seen from the doc link you provided, it seems that you are using Nebula Graph v1.x. To clarify, rank is a value to differentiate the multiple edges of the SAME edge type between TWO vertices, i.e. the same starting vertex and destination vertex. The results of your query are different destination vertices. So it is not following the rule of descending order. In your case, if there are five edges of the same edge type between "11111111111111111111" and "22222222222222222222" and the edges are assigned different rank values, then these edges should be in descending order in the results.
Does that make sense to you?

#4

You are right - the docs are nebula 1.
Sorry for disturbing.

1 Like