cell values in the logsheets that are 'empty' should be transformed into 'NA'

[laurianvm]

@isanti and @cpavloud can you comment on this: empty cells in the logsheets should be tranformed into NA (or none) in the "transformed" logsheets because there should never be an empty cell (rules of interoperable and reusable spreadsheets).
BUT do we want that this value of NA (or none) is also copied over to the triples, which will become our triple store and what the VRE will access, for those columns that are mandatory. (for the non-mandatory columns then we would simply not have any entry at all).
I think we did talk about this ages ago, but I dont remember the answer. I have a feeling the decision was to not include any statements in the triples if there is no data in the logsheet, for the mandatory as well as the non-mandatory columns.

[isanti]

empty cells in the logsheets should be tranformed into NA (or none) in the "transformed" logsheets because there should never be an empty cell (rules of interoperable and reusable spreadsheets).

Yes, as described here, for interoperable and reusability.

do we want that this value of NA (or none) is also copied over to the triples, which will become our triple store and what the VRE will access, for those columns that are mandatory.

When there is an NA the user (human/machine) will know instantly there is nothing to expect here. If they don't find a mandatory column they might start looking for it. I would say that NA are added in the mandatory columns because there is a meaning to this NA (something went wrong, was not recorded etc).

for the non-mandatory columns then we would simply not have any entry at all

yes

[bulricht]

Just for clarification: a triplestore doesn't store data as a table, but as a graph. By consequence, the concept of "columns" does not exist there. Additionally, for tabular data, I would advise against persisting NA into the data. NA would just be another name for NULL, it would mean that mandatory is not actually mandatory and all software working with this data will be prone to errors (data software supports NULL values out-of-the-box, since that is a standard, but NA values need to be handled manually each time).

[isanti]

ok. I see thanks for explaining this.
Then in that case, I understand it makes no sense to have NULL data branches in the graph. And, it would also create a lot of manual work. Let's proceed like this and have the NA columns (mandatory or not) without any entry in the graph.