Hi,
I noticed an inconsistent behavior of the Reference Column Resorter compared to the corresponding Column Filter and Splitter Nodes.
The Column Reference Filter Nodes can cope with an empty input table as long as it has columns. However, the Reference Column Resorter fails to execute.
Reference Column Resorter 4:10 Order column "null" not present in 2nd input table
I assume this is likely related to the fact that the Reference Column Resorter Node is using Column Names as Strings per row instead of a reference table with the column names (and types) like the Reference Column Filter and Splitter do.
Shouldn’t the structure of the input data for the same functional group of nodes be identical or at least similar to ease interaction and not having to think (manage inconsistencies instead of focusing on the primary objective)?
Best
Mike