Frequently Asked Questions
If you upload the blob before the object, is there a risk of orphaned blobs?
Yes. Blobs can end up being orphaned, if the client didn't complete the object creation step. This is preferable to having an object being uploaded with links to blobs that do not exist. We upload data first and then the object from service.
Can we work with parts of a huge object or do we need to load all of it and then work with it?
Yes. Parquet allows access to row-groups without having to read the entire blob.
Are data blobs somehow linked to the object(s)?
Objects can contain a data_id. This identifier is the hash of a data blob. The identifier allows us to see what data is related to an object, but not the other way around.
What happens when I want to rename an object?
If an object in app is renamed (e.g. in the Project Tree), the object will not be renamed in Seequent Evo. This is because V1 Object Service does not yet provide rename functionality. The object in app will be linked to object in Seequent Evo Storage, so it is recommended that you can find the related path, with published object name, so that it is known which object is linked.
Can I export my object?
Export is done in app only — it is not currently a feature of the connected workflow.
Can the API handle Cyrillic/other UTF-8 characters?
The object identifiers must be ASCII letters, numbers, or any of these special characters: +-.:=_/.
Will data be dynamically updated when the schemas are updated?
A new version of the object has to be uploaded to update the version of a schema used by an object.
Where are the data validations done?
Data validation is done by the schema — the API does not validate data. Object API validates the schema.
Can we share a object via a URL or by some other means?
Objects will have a UUID that can be shared, but no shareable link is currently available.
Is there any way to retrieve the UUID of an outdated object?
You can use the list endpoint for your data location and retrieve the UUID for it.
