Is it possible to apply `by` modulator to not last step?
I have
TheCrew
graph.
I'm trying to get 2 properties and unfold values only for second one, query is like
g.V().valueMap("location", "name").by(__.unfold()).by()
but property order after valueMap
can be random, so I added sorting to get expected result
g.V().valueMap("location", "name").order(local).by(keys)
Is it possible to add by
modulator after sorting to valueMap
maps?
g.V().valueMap("location", "name").order(local).by(keys).by(__.unfold()).by()
get error because try to apply all by
to order
step5 Replies
property order after valueMap can be random
by()
is order dependent to the specification of the property keys in valueMap()
. the first by()
must always apply to "location" and the second by()
must always apply to the "name".gremlin> g.V(1).valueMap("location", "name").by(__.unfold()).by()
==>[name:marko,location:[san diego,santa cruz,brussels,santa fe]]
strange. i feel like something must be broken
wonder how long that's been misbehaving
tests must have been passing by accident all this time. or do we not test that?
Current implementation return properties in order by addition.
Cole's shuffle tests found this issue
hmmm - i wonder if graphs are returning properties in the way this shuffle is forcing?
i'd assume we'd have run into this before otherwise.
it seems that there is a bug in this behavior https://issues.apache.org/jira/browse/TINKERPOP-2974