Incorrect relation matching when using custom ID field

API & Programming
1

Hi there,

I’m experiencing an issue with the Fibery Integration API where entities are being linked incorrectly despite sending correct data. Not sure if this is an API issue or if I just didn’t quite understand how the intended usage is.

Setup:

  • We’re syncing data from an external system (Jira Assets) to Fibery
  • Using the /api/v1/synchronizer/data endpoint with synchronizationType: 'full'
  • Entities have a custom id field (using keys like “ES-123”, “ES-456”) instead of UUIDs
  • Relations are configured with targetFieldId: 'id' to match against this custom field

The Problem:
When syncing a “Measure” entity (let’s call it “M-1312”) that should only be linked to one “School” entity (“S-49”), Fibery is incorrectly linking it to multiple schools (both “S-4” and “S-49”).

What we’re sending:

{
  "id": "M-1312",
  "name": "Measure 1312",
  "schulen": ["S-49"]
}

Target entities in Fibery:

// School S-4
{
  "id": "S-4",
  "name": "School 4"
}

// School S-49  
{
  "id": "S-49",
  "name": "School 49"
}

Expected behavior:
M-1312 should only be linked to S-49 (since schulen only contains ["S-49"] and the relation uses targetFieldId: 'id' for exact matching).

Actual behavior:
M-1312 is linked to BOTH S-4 and S-49.

Schema configuration:

"schulen": {
  "type": "array[text]",
  "relation": {
    "cardinality": "many-to-many",
    "targetType": "school",
    "targetFieldId": "id"
  }
}

Questions:

  1. Does Fibery perform any fuzzy matching on the id field (e.g., prefix matching)?
  2. Should the id field be of a specific format?
  3. Is there a known issue with using non-UUID values in the id field for relation matching?

Any insights would be appreciated!

Alex

2

Fibery uses contains for matching, so indeed, if you are matching on ‘S-49’ then both ‘S-49’ and ‘S-4’ will be linked.
I’d suggest doing some kind of hashing or string wrapping of the values prior to matching, so that you can avoid this.

3

Thanks a lot for your quick reply! I will create a UUID and use that.

From an implementation perspective it really does not matter, but I don’t find it obvious that you use contains to match two IDs. If you have the option, perhaps add a note to your documentation :wink:

Alex