Unfolding superstructure and SDF changes the SDF

[CHarpprecht]

Hi @romainsacchi ,

I have the following issue when folding and unfolding a superstructure DB and its SDF:

• currently, LCA results cannot be calculated with the unfolded superstructure database due to the formatting of the unfolded SDF file, specifically the delimiter. The original SDF file uses semicolons as delimiter, which is great and works well.
• The unfolded SDF uses a mix of comma and " as delimiter. This causes issues, as some of the product names or activity names contain one or more commas, which are then interpreted as delimiter. But the unfolded SDF also uses "" in some cases (I believe in those cases, where additional commas are used in the activity or product name). This mix of delimiters makes the csv not usable, as it is not clear what is a delimiter and what is part of an activity name.

Would it be possible to just switch to semicolons as well for the unfolded SDF?

[CHarpprecht]

@romainsacchi , the initial issue raised might not be problematic as it is just a different (but valid?) formatting.

However, I think there is an issue regarding the actual content of the unfolded SDF:

The Scenario difference file (SDF) unfolded does not seem to be identical to the original SDF.

• Tested with a superstructure for 2 scenarios. Thus the original SDF contains 2 columns with values for those scenarios. Cols are labelled like the original scenario names (SSP2-M, etc.).
• But the unfolded SDF only contains 1 col with values instead of 2 columns. The label of the column is not a scenario label, but the Database name. Where did the other column go and which scenario values are in the unfolded SDF?
• The behaviour I expected would be that the unfolded SDF and SS-DB can be used to produce exactly the same results as with the original SS-DB and original SDF.

[CHarpprecht]

@romainsacchi , the initial issue raised might not be problematic as it is just a different (but valid?) formatting.

However, I think there is an issue regarding the actual content of the unfolded SDF:

The Scenario difference file (SDF) unfolded does not seem to be identical to the original SDF.

• Tested with a superstructure for 2 scenarios. Thus the original SDF contains 2 columns with values for those scenarios. Cols are labelled like the original scenario names (SSP2-M, etc.).
• But the unfolded SDF only contains 1 col with values instead of 2 columns. The label of the column is not a scenario label, but the Database name. Where did the other column go and which scenario values are in the unfolded SDF?
• The behaviour I expected would be that the unfolded SDF and SS-DB can be used to produce exactly the same results as with the original SS-DB and original SDF.

It could be that the original SS-DB and SDF were already not working. They were created with premise version 2.3.4., and somehow the keys in the SDF and the SS-DB seemed to be not matching in the original DB already. Updating to premise version 2.3.6 resolved the issue of unmatched act-keys in the original SS-DB and SDF. But I don't know whether that was the root of the issue leading to an erroneous unfolding of the SS-DB and SDF.