groupColumn cannot be a column added by the mapper

[nkuehn]

Hi,
I just tried a mapping that uses a column that's added by the same mapping file as "groupColumn".
Result is an empty CSV with just the column names line.

It would be nice if this were supported or at leas explicitly caught and/or documented.

[OlegIlyenko]

groupColumn was designed to preserve the order of rows from the original file (or at least subsets of the rows). Can you please explain you use-case? It would be really helpful, especially how you create this new column: is it copied from some other column with or without transformations?

[nkuehn]

HI, in this case the grouping "ID" is a concatenation of three other identifiers in the source data (category, Manufacturer, Model), which define a group only "combination unique". To me it seems rather like a question of processing order: Process Column transforms and additions first, then do the grouping, then add the variant counter column. But that may be a limited view of what's going on inside.