Reconcile sensor and level between Kvalobs and LARD

[Lun4m]

A missing (sensor, level) pair in the Obsinn message format is treated differently by Kvalobs and LARD.

Obsinn param code	Kvalobs Sensor Level	LARD Sensor Level
TA	('0', 0)	(NULL, NULL)
TA(0,0)	('0', 0)	(0, 0)

While migrating from Kvalobs, it's impossible to tell the difference between the two cases, unless the timeseries is already present in LARD. And after the migration, we might end up with timeseries data split between two labels.

[ketilt]

Will it help to have a tool that dedupes cases of suspected duplication? That is, if similar timeseries are checked to be fully identical, then the "least preferred" one will be removed and archived?

[Lun4m]

Yes, that's the idea, unless we decide to change how LARD works in this regard