Skip to main content
Gathering Interest

Data Model - Transaction Commit/Rollback Capabilities

Related products:Data Modeling
  • January 5, 2026
  • 3 replies
  • 68 views

Hello,
We wanted to check if we can have transaction capabilities when inserting or deleting data in data model in Cognite using SDK.

Right now we have requirement to insert or delete into multiple views corresponding to single operation on UI Application, so either operation should be marked as done completely and data should be affected in all views or operation should be failed and data should not be affected partially. So either whole transaction is committed or whole is rollbacked.

If it is already available can you share the documentation link.

If not available can we take this as requirement. Please share ETA around the same.

3 replies

Magnus Helgeby
Practitioner ⭐️⭐️⭐️
  • Practitioner ⭐️⭐️⭐️
  • January 5, 2026

Hi ​@Neerajkumar Bhatewara - thank you for your feedback on this topic. I have forwarded your request to the responsible PM who will review your request shortly. 


  • Practitioner ⭐️⭐️⭐️
  • March 10, 2026

Hi ​@Neerajkumar Bhatewara, could you please clarify the concern here a bit more?

Writing to multiple Views in a single transaction/request is an atomic operation, so this scenario you describe should never happen. Deleting an instance deletes all its properties in all containers, so that is also atomic in nature.

Would appreciate any more context here regarding this request.


Regards,

Uzair


!--startfragment>

Thanks for reaching out, ​@Uzair Wali 

As mentioned, a single operation might trigger multiple insertions and modifications across several views. Later, while adding data into the second or third view, the operation may fail for valid reasons. In such cases, we want to revert the insertion or update that happened in the first view during the initial step.

If we perform this as one transaction, we can roll back the entire transaction if the second or third step fails.

Possible failure reasons include CDF connectivity issues, data constraints, and similar factors.

!--endfragment>