Fabric graph

[psrihitha]

Can we load the data into fabric graph from csv using cypher.If so could you please help me

[v-csrikanth]

Hi @psrihitha 

Hi @,

We wanted to follow up and check whether you had a chance to review our previous response and if the issue has been resolved.

If you are still experiencing the issue or have any additional questions, please feel free to share an update. We’ll be happy to assist further.

Thank you,
C Srikanth
Community Support Team

[psrihitha]

Hi @v-csrikanth could you please help me with this.  while creating a node in fabric graph do we have an option of merge smilar to neo4j in fabric

[psrihitha]

Hi @v-csrikanth 

The issue has been resolved.I have combined all the user details and created the user node and mapped the respective columns i.e owned_by,managed by to the user node and  created owns and manages relationship

[v-csrikanth]

Hi @psrihitha 
We would like to inquire whether have you got the chance to check the solutions provided by @rohit1991 in community to resolve the issue. We hope the information provided helps to clear the query. Should you have any further queries, kindly feel free to contact the Microsoft Fabric community.

[psrihitha]

If i had to create user node as below do i have to create 3 different user nodes.And one more doubt is can we create the same node from different tables 

[rohit1991]

Hii @psrihitha 

You do not need to create three separate User node types. Create one User node type with UserID as the unique key; each distinct row/UserID in the source table automatically becomes a separate User node.

A single node type currently maps to one source table. If user details are spread across multiple tables, first merge/union and deduplicate them into one Lakehouse table or view. Different relationship tables can still create multiple edges connected to the same User node by matching the same UserID.

[psrihitha]

In my case, the user-related data is distributed across multiple tables and relationships originate from the same User node.How can i determine which relationship should be created for each record?Data contains Owned by/Managed by/IT Application owner user details.

[rohit1991]

Hii @psrihitha 

Keep one User node using UserID as the key and one Application node using ApplicationID. Create separate edge types for each role: User-OWNS>Application, User-MANAGES>Application, and User-IT_APPLICATION_OWNER_OF>Application. Map each edge to its corresponding column, such as OwnedByUserID, ManagedByUserID, or ITApplicationOwnerUserID; the edge mapping determines the relationship for every record. The same source table can be used for all three edge types when it contains these columns.

When the data has a single RoleType column, either prepare three filtered Lakehouse tables before creating the edges, or create one generic HAS_ROLE edge and add RoleType as an edge property. Fabric creates one edge per source row, and the User/Application IDs and data types must match their respective node keys.

[rohit1991]

Hii @psrihitha 

CSV data can be used, but not loaded directly through Cypher. Upload the CSV to a Fabric Lakehouse, load it as tables, then create a Graph model > Get data, select those tables, and define the nodes, edges, and key mappings. Fabric Graph uses GQL, not Cypher, for querying the graph.

The link below explains this:

https://learn.microsoft.com/en-us/fabric/graph/quickstart