This quote from the FME 2014 documentation suggest that it should handle arbitrary schema:
"... and is very useful for solving problems when the input format and schemas are not known at design time."
However I have a workflow with Generic (Any Format) Readers that throws an error when I change one or more of the input datasets when testing. The error is:
"During translation, some features were read that did not match a reader feature type in the workspace.
This can happen if the reader is changed, or a reader feature type removed or renamed."
"The following unexpected feature types were read:"
.....
My workflow checks for a number of data problems. There are two datasets: a reference dataset and a dataset to check. The reference dataset and the one to check are each used by a Schema (Any Format) Reader so I can check schema differences and a Generic (Any Format) Reader for finding attribute and geometry problems.
I am using FME 2014 build 14290 on Windows 7 64bit.
Philip