How to get edge properties in query result?

Neo4j Graph Platform
1

How do I get properties of edges and not just nodes from a query to the python driver?

I have a query to try and find variable length paths between two nodes, like

match path = (n1:page{name:'start-page'})-[*..2]->(n2:page) return path limit 5

In the neo4j browser, table view I can see a table with a segments property in the middle with all the data on each connecting edges (see below)

But when I send the same query to cypher/python driver, I don't get any segments, just the start/end nodes.

What query can I use to get the same detail on the path / edges as below?
Do I need to use subgraphs or somehow unwind the paths between?

{
  "start": {
   "identity": 95, "labels": ["page"],
 ...
  "segments": [
    {
      "start": {
         "identity": 95,
         "relationship": {
"identity": 152,
"start": 95,
"end": 97,
"type": "nav",
"properties": {
"start_page": "start-page",
"weight": 2,
...
  "length": 1.0
}

I did try a more complex with type query, which works well in neo4j browser, but from the python driver I still just get the pages, and no data or properties on the edges between each node:

// unex charge to end by weight
match path = (n1:page{name:'start-page'})-[*..5]->(n2:page{name:'end_session'}) 
with n1 as startPage,
    relationships(path) as edges,
    n2 as endPage
return startPage, edges, endPage 
order by edges[0].weight desc
limit 50

This just seems to put page data into the edges too:

 {
 	'edges': [({
 			'flow': 'BILL',
 			'name': 'start-page',
 			'type': 'page',
 			'weight': 18
 		},
 		'nav', {
 			'flow': 'BILL',
 			'name': 'end_session',
 			'type': 'page'
 		})
    ],
 	'endPage': {
 		'flow': 'BILL',
 		'name': 'end_session',
 		'type': 'page'
 	},
 	'startPage': {
 		'flow': 'BILL',
 		'name': 'start-page',
 		'type': 'page',
 		'weight': 18
 	}
 }

Confused why I can't get edge data, and also what magic the neo4j query browser is doing to get queries to work, in a completely different way than the python driver. >.<

2

Hello @D_C :slight_smile:

You should have a look at the properties() function:

MATCH path = (n1:page {name:'start-page'})-[*..5]->(n2:page {name:'end_session'}) 
WITH n1 AS startPage, [p IN relationships(path) | properties(p)] AS edges, n2 AS endPage
RETURN startPage, edges, endPage 
ORDER BY edges[0].weight DESC
LIMIT 50

Regards,
Cobra

3

Thanks @Cobra actually didn't know about this. That does get a bit further.. it allows me to get the list of edges data, but not any data on in-between nodes.
eg I can get

{startNode},
[ {edge}, {edge}, {edge} ... ]
{ endNode }

The benefit of the patterns like -[*..2]- is that I can have a varying number of edges AND nodes.
What I really want is just a sequence of objects, of varying types, eg:

[ 
  [ {node}-{edge}-{node} ],  // first path
  [ {node}-{edge}-{node}-{edge}-{node} ] // longer path
]

Which is exactly what I see in the neo browser in the table view as 'segments'

I'm confused why cypher queries return one type of data structure in the neo-browser, but completely different and less useful data when run from the python driver?

Do I need to use subgraphs or something more complex?

4

Hi,
there is an API documentation on paths that you can use to deconstruct them

https://neo4j.com/docs/api/python-driver/current/api.html#path

5

I can't really deconstruct the paths if there's no data there.
There is literally just start and end nodes and nothing in-between when I look at .data().

I did look at .graph but that also seems not to work as expected

The super-frustrating thing is the neo browser shows all the data I want in the table view.
So neo4j core DB must be capable of giving the data...

  • Is this a case of "private APIs" that return data via.. maybe bolt protocol?

  • Are the cypher queries that I type into the neo browser somehow modified/augmented to get useful data?

  • does the official python lib not support paths?

  • How about the node drivers, do they behave closer to expectation?

neo4j works well as a kind of lightweight data exploration app but using those same queries via the API is another matter.

6

I don't know me being a python-noob I didn't have problems following the API docs to come up with this:

https://neo4j.com/docs/api/python-driver/current/api.html#graph-data-types

from neo4j import GraphDatabase,basic_auth

driver = GraphDatabase.driver("bolt://localhost",auth=basic_auth("neo4j","test"))

cypher_query = "MATCH p=(n)-[r]->(m) RETURN p LIMIT 5"

with driver.session(database="neo4j") as session:
  g = session.read_transaction(
    lambda tx: tx.run(cypher_query,).graph())

  for n in g.nodes:
    print("id %s labels %s props %s" % (n.id, n.labels, n.items()))

  for r in g.relationships:
    print("id %s type %s start %s end %s props %s" % (r.id, r.type, r.start_node.id, r.end_node.id, r.items()))

output

id 10 labels frozenset({'LittleSisEntity', 'Person'}) props dict_items([('littleSisId', '102193'), ('name', 'DONALD J. TRUMP'), ('pageRank', 0.15000000000000002), ('wcc_partition', 9)])
id 473 labels frozenset({'LittleSisEntity', 'Person'}) props dict_items([('littleSisId', '13443'), ('name', 'JOHN SIDNEY MCCAIN'), ('pageRank', 0.15000000000000002), ('wcc_partition', 472)])
id 41 labels frozenset({'LittleSisEntity', 'Person'}) props dict_items([('littleSisId', '13191'), ('name', 'HILLARY DIANE RODHAM'), ('pageRank', 0.15000000000000002), ('wcc_partition', 40)])
id 2388 labels frozenset({'LittleSisEntity', 'Person'}) props dict_items([('littleSisId', '34136'), ('name', 'WILLARD MILTON ROMNEY'), ('pageRank', 0.15000000000000002), ('wcc_partition', 2387)])
id 15368 labels frozenset({'LittleSisEntity', 'Person'}) props dict_items([('littleSisId', '13536'), ('name', 'CHARLES B. RANGEL'), ('pageRank', 0.15000000000000002), ('wcc_partition', 15367)])
id 16822 labels frozenset({'Organization', 'LittleSisEntity'}) props dict_items([('littleSisId', '28778'), ('name', 'REPUBLICAN NATIONAL COMMITTEE')])
id 47647 type DONATION start 10 end 473 props dict_items([('relDetail', 'Campaign Contribution'), ('source', 'LittleSis')])
id 72375 type DONATION start 10 end 41 props dict_items([('relDetail', 'Campaign Contribution'), ('source', 'LittleSis')])
id 72385 type DONATION start 10 end 2388 props dict_items([('relDetail', 'Campaign Contribution'), ('source', 'LittleSis')])
id 72377 type DONATION start 10 end 15368 props dict_items([('relDetail', 'Campaign Contribution'), ('source', 'LittleSis')])
id 72384 type DONATION start 10 end 16822 props dict_items([('relDetail', 'Campaign Contribution'), ('source', 'LittleSis')])