Hi, i found a possible bug/ problem

Basically have to merge 2 datacubes with the same extent, each of them contains the same 2 bands named " b02" and “b03” that i get form some process and one date ( the dates are different )

I did not specify a resolver because i want to keep the 2 bands for each date

if cube is None:
    cube = str_bands_date
    cube = cube.merge_cubes(str_bands_date)       

When i try to execute the process:

[500] OverlapResolverMissing: merge_cubes: Overlapping data cubes, but no overlap resolver has been specified. Either set an overlaps_resolver or rename the bands. Left names: [‘B02’, ‘B03’], right names: [‘B02’, ‘B03’]. (ref: r-24032804d38344cb95cdf585efebbd0c)

Am i doing something wrong ? can you help ?


Hi Nicola,
the solution is to provide an overlaps resolver, something like a ‘max’ process or whatever.
If the dates are indeed different, the resolver will not have to be used in practice, and you’ll get the desired merged cube.

best regards,

Great ! Now It works, very helpful and responsive as always !

Best regards , Nicola

@jeroen.dries thanks for the clarification. From my point of view the user shouldn’t be forced to select an overlap_resolver in this scenario, even if in the process definition this case is not explicitly covered: openEO API Processes

FYI: this is a know issue: merge_cubes: only require overlap resolver when there is actual overlap · Issue #479 · Open-EO/openeo-geopyspark-driver · GitHub