Node type of shortest path

Neo4j Graph Platform Cypher
1

Hi,

Newbie here, please be gentle.
Background mapping patents and node type include
patent
company (who is the patent assigned to)
classification (how was the patent classified) There are actual 4 types of classification nodes
group
subgroup
section
subsection

Testing shortest path with the following Query:

MATCH (source:company{name:'Applied Materials'}),
(destination:company{name:'Sharp Kabushiki Kaisha'})
CALL algo.shortestPath.stream(source,destination,null)
YIELD nodeId
RETURN algo.getNodeById(nodeId).PatNum as PatNum

yields
PatNum
null
"8183081"
null
"8258596"
null

if I change the lest line to
RETURN algo.getNodeById(nodeId).name as company
gets this

company
"Applied Materials"
null
null
null
"Sharp Kabushiki Kaisha"

RETURN algo.getNodeById(nodeId).subgroup_id as subgroup
gets
subgroup
null
null
"H01L31/02242"
null
null

Is there a Query that would the node type?

such it returns
type
company
patent
subgroup
patent
company

Andy

2

LABELS(node) will give all the labels on the nodes (i.e node type as you call it :slight_smile: ) .
It returns array of Strings .

3

Thanks,

When I run this:
MATCH (source:company{name:'Applied Materials'}),
(destination:company{name:'Sharp Kabushiki Kaisha'})
CALL algo.shortestPath.stream(source,destination,null)
YIELD nodeId
RETURN Labels(algo.getNodeById(nodeId))[0] as type

I get which is basically what I want.
type
"company"
"patent"
"Subgroup"
"patent"
"company"

And follow on question: How do I then return a specific property based on the node type? If it is a "company" return name, "patent" return patnum, and subgroup return subgroup_id?

I would like to end up with a table like this.
type value
"company" "Applied Materials"
"patent" "818308"
"subgroup" "H01L31/02242"
"patent" "8258596"
company Sharp Kabushiki Kaisha

Andy

4

You can't do that automatically, though you can do that explicitly using CASE. kind of like:

WITH algo.getNodeById(nodeId) as node
WITH node, Labels(node)[0] as type
RETURN type, CASE type WHEN "company" THEN node.name WHEN "patent" THEN node.patnum WHEN "subgroup" THEN node.subgroup_id END as value

Alternately you could use coalesce() to supply those properties, and the first non-null access will be used (this won't be a good idea if some of those properties are also present on the other nodes).

WITH algo.getNodeById(nodeId) as node
WITH node, Labels(node)[0] as type
RETURN type, coalesce(node.subgroup_id, node.patnum, node.name) as value
5

Hi Andrew,

Thank you very much.

Question for my education:

in this line
WITH node, Labels(node)[0] as type

why is "node," needed? are each WITH statements standalone clauses? Is was defined in the line above.
Andy

6

WITH clauses limit which variables are in scope, so if I hadn't included node in the clause along with the projection of type, then node would be out of scope and no longer usable in the rest of the query.

7

Thanks for the clarification.

I suggest that the documentation could make this point a bit more explicitly.
Andy

1 Like
8

It's a good suggestion, I've reviewed the docs over the last couple versions and agree. We'll push for a an update to clarify this.

1 Like