I'm reading a feature layer from ArcGIS Online containing about 20 layers into FME, each one with proper domains and field aliases. Unfortunately, it is not possible to: first, resolve domains without including only code output attributes, and second, there is no option to "resolve" Attribute names using aliases.

I'm then trying to find a time-efficient way to solve this by using AttributeRenamer. I'm currently importing an Excel schema that contains only two columns for Input and Output Attribute and the list of attributes I want to keep from the reader output. In the Import Wizard, I deselect "Sorted" (img 1) to keep the original order of attributes and then I set the "If Input Attribute Is Missing" to "Remove Output Attribute" (img 2).

What I'm experiencing is that imported attributes from the schema are still alphabetically sorted despite the "Sorted" box being deselected (img 4) and second, when I run the transformer, I can see that all attributes that were not in the Excel schema have been appended at the end of the table, instead of being removed (img 3).

Is there something I'm not doing correctly?