On the reader featuretype you can expose the gml_id and then use an AttributeCopier to copy the ID into an attribute with a name of your choosing. You can also use an AttributeExposer instead of exposing the attribute at source.
If there are many parts to each feature you may first need to use a Deaggregator to 'explode' each building into its component parts, though take care as you may have aggregates inside aggregates in which case you may need to 'flatten' more than one level.
On the reader featuretype you can expose the gml_id and then use an AttributeCopier to copy the ID into an attribute with a name of your choosing. You can also use an AttributeExposer instead of exposing the attribute at source.
If there are many parts to each feature you may first need to use a Deaggregator to 'explode' each building into its component parts, though take care as you may have aggregates inside aggregates in which case you may need to 'flatten' more than one level.
Wow, that was fast. Thanks much! AttributeExposer did the trick.