Neptune Query Returning Duplicate objects
This is how my data in graph db looks. There are three types of labels
- Part
- SubstitutePart
- Tool
And two types of edges
- Substitute: Part -> SubstitutePart
- Uses: SubstitutePart -> Tool Or Part -> Tool
I am trying to fetch all the vertices with label as 'Tool' whose SubstitutePart Vertex Id is "someId" and I am also traversing to the corresponding Part vertex to get the all the Tool vertices from there as well. So, In-short all the Tool vertices that are directly or indirectly related to a specific SubstitutePart vertex.
My gremlin query that I POST to Neptune db using REST API looks like this:
{
"gremlin": "g.V('someId').hasLabel('SubstitutePart').as('subPart').outE('uses').inV().as('subTools').in('uses').inE('substitute').outV().outE('uses').inV().as('partTools')"
}
And the response I get from this query is this.
{
"requestId": "1d256612-1716-8457-b8ac-2442666d174c",
"status": {
"message": "",
"code": 200,
"attributes": {
"@type": "g:Map",
"@value": []
}
},
"result": {
"data": {
"@type": "g:List",
"@value": [
{
"@type": "g:Vertex",
"@value": {
"id": "1c256612-1716-4b76-b8ac-2472666d17ec",
"label": "Tool",
"properties": {
"number": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": 678255301
},
"value": "2",
"label": "number"
}
}
],
"versionId": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -1536929837
},
"value": "1c256612-1716-4b76-b8ac-2472666d17ec",
"label": "versionId"
}
}
],
"creationDate": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -1328847462
},
"value": {
"@type": "g:Int64",
"@value": 0
},
"label": "creationDate"
}
}
]
}
}
},
{
"@type": "g:Vertex",
"@value": {
"id": "1c256612-1716-4b76-b8ac-2472666d17ec",
"label": "Tool",
"properties": {
"number": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": 678255301
},
"value": "2",
"label": "number"
}
}
],
"versionId": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -1536929837
},
"value": "1c256612-1716-4b76-b8ac-2472666d17ec",
"label": "versionId"
}
}
],
"creationDate": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -1328847462
},
"value": {
"@type": "g:Int64",
"@value": 0
},
"label": "creationDate"
}
}
]
}
}
},
{
"@type": "g:Vertex",
"@value": {
"id": "f38e06dc-b42f-4236-a260-082c53b4d09c",
"label": "Tool",
"properties": {
"number": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -24263995
},
"value": "1",
"label": "number"
}
}
],
"versionId": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": 1752257459
},
"value": "f38e06dc-b42f-4236-a260-082c53b4d09c",
"label": "versionId"
}
}
],
"creationDate": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -1498607962
},
"value": {
"@type": "g:Int64",
"@value": 638125689707921052
},
"label": "creationDate"
}
}
]
}
}
},
{
"@type": "g:Vertex",
"@value": {
"id": "f38e06dc-b42f-4236-a260-082c53b4d09c",
"label": "Tool",
"properties": {
"number": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -24263995
},
"value": "1",
"label": "number"
}
}
],
"versionId": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": 1752257459
},
"value": "f38e06dc-b42f-4236-a260-082c53b4d09c",
"label": "versionId"
}
}
],
"creationDate": [
{
"@type": "g:VertexProperty",
"@value": {
"id": {
"@type": "g:Int32",
"@value": -1498607962
},
"value": {
"@type": "g:Int64",
"@value": 638125689707921052
},
"label": "creationDate"
}
}
]
}
}
}
]
},
"meta": {
"@type": "g:Map",
"@value": []
}
}
}
The result has 4 vertices and 2 of them are duplicates. It is containing the Tool vertices which are directly linked with Part label. Vertices that are linked with SubstitutePart labels are not returned in the response.
What should be fixed in the query in order to get all Tool vertices that are directly/indirectly linked with SubstitutePart label vertex?
It appears like you're walking the graph and then overlapping the path from which you came. What you really want is a Gremlin query that will spawn traversers [2] and walk both parts of the graph "simultaneously" (or, at least appear as if it is happening in parallel).
Instead, you may want to do:
g.V('someId').
//hasLabel('SubstitutePart'). [1]
union(
out('uses'),
in('substitute').out('uses')
)
The union will spawn two traversers to "walk" the graph in the opposite directions.
It also appears that you're using as() labels within your query. These are only needed if you then want to use something like a select() later in the query to fetch whatever items you "labeled" using that as() label.
[1] There's no need to check for a Label if you're looking up a vertex by an ID. IDs in TinkerPop are unique per every vertex or edge (same goes for Neptune).
[2] Think of a traverser as a little Gremlin that spawns every time there is a fork in the graph that needs to be traversed. In actuality, this is the query engine keeping track of where in the graph we've been and where we are headed.
- Little hack for ggplot -- an easy way to add a text with the real means and standard deviation when using lines or bars
- graph_to_gdfs not working even with simple examples
- Single Source - Single Destination Shortest Path (DFS + DP) - Time complexity?
- get all possible paths in directed cyclic graph
- How can I see all the SharePoint Sites a user is an owner of with Microsoft Graph?
- how to use vector to store a shape?
- Clear and automatic DAG visualization
- What's the difference between undirected graph and unweighted graph? Are they the same thing?
- Draw a horizontal and vertical line on mouse hover in chart js
- Linear time algorithm for computing radius of membership hyper-sphere
- Why is BFS is much quicker than DFS when I implement both of them?
- Command-line Unix ASCII-based charting / plotting tool
- R ethogram or actogram script
- ggarange: graphs with different size legend
- Replacing all relationships between nodes in Neo4j
- Topological sort not printing all vertexes
- How to replace NULL in dataframe with 0?
- Excel graph date and count
- How to add a feature to all the nodes in networkx
- How can I make a 3D bar graph and plot it on excel using data where the negative values on the graph will be colored as well?
- Travel from city 1 to city n, visiting at least 3 odd cities
- Inconsistent graphs with Altair despite same package version
- Pyvis network keeps on moving
- Applications of Voronoi without Delaunay?
- How do I correctly draw edge between records with Graphviz and Haskell
- Pyside6: Dynamically add points to chart view
- Kruskal's algorithm including a specific edge
- Using networkx to create node labels from an array
- Topological sort on directed and undirected graphs using DFS algorithm
- Adding a regression plane to plot_ly 3d plot