Asset import with reference object

[Ismael]

Hi everyone,

I'm importing objects of the type Voice Service from a CSV file into Jira Assets. One of the columns (F) contains references to objects of the type Voice Technical Solution. These references are stored using an identifier from the "Voice Technical Solution" object, and when there are multiple references, they are separated by ||.

The attribute in Jira Assets is configured as an object reference, and I'm using the following AQL during import:

objectType = "Voice Technical Solution" AND "Identifier" IN "1"

The attribute Identifier is a text field. However, even when a row contains only one reference, the relationship is not created.

I’ve confirmed that the referenced objects exist and that the attribute is correctly defined. Has anyone encountered this issue or knows how to resolve it?

Thanks in advance!

[Jason U]

This usually happens because the AQL reference doesn’t match exactly or because multiple reference values in the CSV aren’t supported during import.

Try testing with a single reference value (no “||”), and confirm the attribute mapping uses AQL match.

If the single-value import works, you’ll need to add the multiple references later using a bulk update or the Assets REST API.

Reference: Atlassian Developer Docs – Insight REST API

[Ismael]

With identifier = $(voice_technical_solution), the CSV only loads objects that have a single object relationship, not those with multiple ones. It does perform searches with Identifier = "Instituto de Málaga". I’ve tried loading the CSV using concatenators like ||, "..."||"..."||"...", and "...","...","...", but it only loads those with a single relationship.

[Jason U]

This behavior happens because Jira Assets (formerly Insight) doesn’t support importing multi-value object references through CSV.
When you use a delimiter like || or ,, the importer only reads the first reference and ignores the rest.

There are two reliable ways to handle this:

Option A — One row per relationship (recommended)

Reshape your CSV so that each relationship between Voice Service and Voice Technical Solution appears on a separate row.
Then, during import, use merge and append settings so multiple rows build up all references for the same object.

Example CSV:

externalId,serviceName,vtsIdentifier

VS-001,Alpha Service,1

VS-001,Alpha Service,7

VS-001,Alpha Service,12

VS-002,Beta Service,5

Import setup:

1. Map externalId as the object key or identifier.
2. Map vtsIdentifier to your reference attribute and match by AQL:
3. objectType = "Voice Technical Solution" AND "Identifier" = "${vtsIdentifier}"
4. Enable Merge with existing objects and choose Append values for the reference field.

Each import row adds one relationship; after import, every Voice Service object will show all linked Voice Technical Solutions.

Option B — Two-phase import with automation or REST API

1. First import the Voice Service objects without the reference, but keep all related IDs in a text field (for example: vtsIdsRaw = "1||7||12").
2. Then use either:
   • Automation for Assets – a scheduled rule that splits vtsIdsRaw, finds each target object via AQL, and appends it to the reference attribute.
   • Assets REST API – run PUT /jsm/assets/workspace/{workspaceId}/v1/object/{objectId}/attributes/{attributeId} for each reference you want to add.

Quick checks

• Make sure “Identifier” on Voice Technical Solution is unique and a text field.
• Always use AQL matching (not name/key).
• Use quotes for strings, e.g. "Identifier" = "${vtsIdentifier}".

[Ismael]

Identifier = ${voice_technical_solution} is the correct way, but with objects with multiple relationships in the CSV it is not support.