One for the engineering team!
I have noticed that when using a NodeByLabelScan - even with a filter - the nodes are always returned by default in strict ascending id order, and with no gaps - that is I don't need to sort by id to get a sorted list of node ids. I'd like to know if this is a guaranteed default behaviour, or is it just a a co-incidence? As illustration of what I mean, these two queries return identical results.
MATCH (n:T) WHERE id(n) >= 100 RETURN id(n) ORDER by id(n) LIMIT 10
MATCH (n:T) WHERE id(n) >= 100 RETURN id(n) LIMIT 10