Closed jnm closed 3 months ago
We could enforce xml_hash
uniqueness before removing the formatting restrictions, but we could not (I was mistaken) enforce UUID uniqueness while the formatting restrictions are in place. That's because we need to identify cases where UUIDs collide but XML content is not identical, and then rewrite their UUIDs to append something like "DUPLICATE n" (which would violate the existing formatting restrictions).
Some internal discussion at https://chat.kobotoolbox.org/#narrow/stream/4-Kobo-Dev/topic/Duplicated.20submission.20UUIDs/near/192516
Closed because it is gonna be handled by kobotoolbox/kpi#5047
The OpenRosa spec allows people to implement a "custom ID scheme" for
<instanceID>
, although it "must be a universally unique string identifying this specific submission".We don't currently support this: https://github.com/kobotoolbox/kobocat/blob/a8fbca5b0f28491de1b4d732bc6ed2339cab0dd6/onadata/apps/logger/xform_instance_parser.py#L72-L93
Some people have uploaded submissions containing the likes of
This results in the
logger_instance.uuid
column diverging from the<instanceID>
in the XML.We should:
logger_instance
table;<instanceID>
string