Currently the schema validation checks .data exists but not it's contents for INamedFile and INamedImage.
But since this is implemented as a property decorator this validation will load the whole file into memory. If this is a large file or you are doing a big import, this can slow down the process considerably as often the data is transferred using fs cp.

I have a PR that avoids this check. #166

This might also be able to be fixed in zope.schema itself - zopefoundation/zope.schema#127

In addition for images the file is being read for unnessarily for exif and content-type extraction. The attached PR also fixes this.