Use Current Column Name in Column Patterns' Value Expression
In any transformation that supports column patterns, it's not possible to reference other columns in the "value expression" based on the name of the current column. This capability would be extremely useful.
I do know about the metadata functions byName, byPosition, ... - none of those are (singularly) appropriate for this scenario.
An example scenario is transforming a dataset that has similarly named columns - such as transactions of different currencies, where the currency values are stored in columns named "Amount ()" where "" is the currency code. The number and naming of the current set of columns is known, but either that number is large (many currencies) or has a possibility of changing (drift) over time, I wish to create a flexible transformation using a column pattern. If I want to create a set of columns based on those "Amount ()" columns named "Formatted Amount ()" I can not do that.
What I would like to be able to do in the "value expression" is something like:
"$" + toString(byName("Amount (" + left(right(name, 4), 3)) + ")")
Ignore that I'm not actually accomplishing much, or even the correct thing with that expression. What I am trying to do is use the "current column name" (say "Amount (USD)") to create a column "Formatted Amount (USD)" from the value that's IN the "Amount (USD)" column.
"$$" and "name" do not function in the "value expression".