Identify similar lines between two networks

I would try buffering your road lines to create areas, then using these areas to clip the paths. Deaggregate the output and then keep only lines that are a significant length. What you should be left with are path segments that are running along/alongside the roads.

I wouldn't expect a perfect outcome from the above, looking at the sample data, but should be a start. Will probably require some trial and error with how much to buffer by.

@djmcdermott

Chop the lines of the datasets by 2. Extract ortientation.

Neighborfinder or Buffer at relevant tolerance.

Compare orientation. Keep those within some tolerance.

This works pretty well. There are a few false positives and not all the lines get line_id's but it works.