This is how it works unfortunately. The Schema is pretty important in FME. You can look into dynamic workflows but that is mostly when you don't know/care what the schema is, you just want to preserve it.
Perhaps what might be useful in your case is to instead read the data based on the position of the data rather than the column names.
You can rename column 'A', 'B', etc to be what you like. In this case the input attributes will always have the same names and you won't need to update the attribute manager.
You will need to remove the header row though.
This method assumes that you will always have the same columns in the same position just sometimes with different names.
This is how it works unfortunately. The Schema is pretty important in FME. You can look into dynamic workflows but that is mostly when you don't know/care what the schema is, you just want to preserve it.
Perhaps what might be useful in your case is to instead read the data based on the position of the data rather than the column names.
You can rename column 'A', 'B', etc to be what you like. In this case the input attributes will always have the same names and you won't need to update the attribute manager.
You will need to remove the header row though.
This method assumes that you will always have the same columns in the same position just sometimes with different names.
Thanks @virtualcitymatt for the info.
