Does it close #501 too? I'm trying to understand, what is the difference from the technical point of view... Like if we can detect bidirectional edges, we can detect undirected too by symmetrize the graph, can't we?

@SemyonSinchenko Thanks for clarification. It is undirected edge. I was trying to include in-bound, out-bound edges.

Bidirectional edge is different concept, it should return (0),(1) only which is not this case. I've changed the word to fix this.

I've changed the title and the code, but cannot change the branch name. It closes the PR. :(

I've changed the title and the code, but cannot change the branch name. It closes the PR. :(

It is not a problem. I linked the right issue.

I am not sure whether we need undirected edge for a variable length pattern (u)-[e*1..3]-(v) or not. I've modified the code though, the result as a tabular format is quite confusing. It is due to the column order changes. Half of them starting from the vertex 2, half of them ending with the same vertex.

+---------+---------------+---------+--------------+---------+--------------+---------+---------------+
|        u|            _e1|      _v1|           _e2|      _v2|           _e3|        v|              e|
+---------+---------------+---------+--------------+---------+--------------+---------+---------------+
|{2, c, m}|{2, 0, unknown}|{0, a, f}|{0, 1, friend}|{1, b, m}|{1, 2, friend}|{2, c, m}|           NULL|
|{2, c, m}|{2, 0, unknown}|{0, a, f}|{0, 1, friend}|{1, b, m}|{1, 0, follow}|{0, a, f}|           NULL|
|{2, c, m}|{2, 0, unknown}|{0, a, f}|{0, 1, friend}|     NULL|          NULL|{1, b, m}|           NULL|
|{2, c, m}|           NULL|     NULL|          NULL|     NULL|          NULL|{3, d, f}| {2, 3, follow}|
|{2, c, m}|           NULL|     NULL|          NULL|     NULL|          NULL|{0, a, f}|{2, 0, unknown}|
|{2, c, m}| {1, 0, follow}|{0, a, f}|{0, 1, friend}|{1, b, m}|{1, 2, friend}|{1, b, m}|           NULL|
|{2, c, m}|{2, 0, unknown}|{0, a, f}|{0, 1, friend}|{1, b, m}|{1, 2, friend}|{2, c, m}|           NULL|
|{2, c, m}| {0, 1, friend}|{1, b, m}|{1, 2, friend}|     NULL|          NULL|{0, a, f}|           NULL|
|{2, c, m}|           NULL|     NULL|          NULL|     NULL|          NULL|{1, b, m}| {1, 2, friend}|
+---------+---------------+---------+--------------+---------+--------------+---------+---------------+

Hi @rjurney, I am reaching out to you again as you opened the feature request #501. Is the behavior what you were looking for? The idea is pretty simple, converting into two part (u)-[e]->(v) and (v)-[e]->(u) and then union.

@rjurney I want to see this feature in a new release and from my point of view, it is a time to start preparing it. What do you think?

I'm merging it now to make it a part of the new release.

@SemyonSinchenko, what is the timeline for the next release? I hoped I could add informational columns _hop, _pattern, _direction. Can I open a new PR for the update?

@goungoun I think we can wait. I was hoping to deliver a new release this week, but we can actually postpone it to the beginning of the next one.