Do Input / Output IDs need to be unique?

[Brilator]

The original annotation principles (and I guess ISA) require that, Inputs and Outputs hold unique IDs.

....“Source Name”, “Sample Name”, “Material”, or “Data”. This input value must be a unique identifier for an organism or a sample.

...Output column is allowed, which can either be a “Sample Name”, “Material”, or “Data”. ... The value of this column must be a unique identifier.

I guess that's not how we use it anymore, right? Otherwise this would disallow to split / pool samples or gain multiple data outputs / merge data inputs from single data files.

@HLWeil @kMutagene @muehlhaus

[kMutagene]

IIRC this means that the identifier must be unique in regards to the entity they describe, not on a column-basis. plain ISA supports pooling/splitting of samples as well: https://isa-specs.readthedocs.io/en/latest/isatab.html#study-table-file

[Brilator]

Ah, alright.