Connected Component Finding

Introduction

Connected Component Finding is an algorithm that identifies all nodes (vertices) reachable from specified starting nodes (vertices) through given edge types. The algorithm calculates the minimum distance (hops) from any start node (vertex) to each reachable node.

Output

The algorithm produces the following fields:

Field Name	Description
ID	The ID of the node (vertex)
hop	The minimum hops from any of the provided nodes (vertices)

Query Example

cyphergremlin

CALL algo.paral.connectedComponentFinding({
    relationshipTypes: ['Knows', 'WorkAt'],
    startIds: ['User[2]'],
    maxIterations: 10
    }) YIELD ID, hop
RETURN ID, hop 

graph.program(
    ConnectedComponentFindingProgram.builder()
    .edges('Knows', 'WorkAt')
    .startIds('User[2]')
    .maxIteration(10)
    .vertexFilter('Company', 'city', TextP.startsWith('B'))                             // filter for one label
    .edgeFilter('Knows', 'startYear', P.gte(2020))
    .vertexFilter('isOutdated', P.eq(false))                                              // filter for all labels
    .edgeFilter('isDeleted', P.eq(false))
    .create()
).submitAndGet().toList()

Parameters

cyphergremlin

Name	Description	Required
relationshipTypes	Relationship types to traverse	Yes
startIds	Initial node (vertex) IDs	Yes
maxIterations	Maximum number of iterations	No

Method	Description	Required
edges(String... names)	Edge types to traverse	Yes
startIds(String... ids)	Initial node (vertex) IDs	Yes
maxIteration(int n)	Sets the maximum number of iterations	No
vertexFilter(String labelName, String Attr, P predicate)	Filter nodes (vertices) of specific label by attribute	No
vertexFilter(String Attr, P predicate)	Filter all nodes (vertices) by attribute (if the node has the given attribute)	No
edgeFilter(String labelName, String Attr, P predicate)	Filter edges of specific label by attribute	No
edgeFilter(String Attr, P predicate)	Filter all edges by attribute (if the edge has the given attribute)	No

cypher is not support filtering nodes and relationships for now.

Exporting Query Results

For large datasets, export the results to object storage:

cyphergremlin

EXPORT TO 's3://my_bucket/target_dir'
PROPERTIES {
  catalog: 'puppygraph_catalog_name'
}
CALL algo.paral.connectedComponentFinding({
    relationshipTypes: ['Knows', 'WorkAt'],
    startIds: ['User[2]']
    }) YIELD ID, hop
RETURN ID, hop 

graph.program(
    ConnectedComponentFindingProgram.builder()
    .edges('Knows', 'WorkAt')
    .startIds('User[2]')
    .vertexFilter('Company', 'city', TextP.startsWith('B'))
    .edgeFilter('Knows', 'startYear', P.gte(2020))
    .vertexFilter('isOutdated', P.eq(false))
    .edgeFilter('isDeleted', P.eq(false))
    .create()
).submitAndSave([
  "exportTo": "s3://my_bucket/target_dir",
  "catalog": "puppygraph_catalog_name"
])