I skimmed over this article:
Based on reading through, there is no way to filter by extension on-import. HOWEVER, what you could do is script a post-import method that goes through, reads the file type, and deletes any records with container fields that contain PDFs.
I believe, based again on what I read, that it does not differentiate PDF files from image files, as you're only setting one of two options ("picture" files or "text" files), since PDFs are not plain text, or have .txt or .rtf extensions, then I would assume PDF falls under the generic "pictures" category.
I believe you can use GetAsText(table::containerField) and it will return to you the file path (if stored as a reference) or filename (if stored internally).
So you could do something like this after your import:
Go To Record/Request - First
exit loop if - get(foundCount) = 0
If ( right ( GetAsText (table::container) ; 4 ) = ".pdf" )
delete record - no dialog
exit loop if - get(recordNumber) = get(foundCount)
go to record - next
someone should check the exit loop if logic above though, it's early in the morning and I haven't had my coffee yet.
Hope this helps a bit.
Thanks for you reply on this issue.
I resorted to a similar method of deleting the PDFs afterwards, until a better method was available. I agree with you, there are no other obvious options when importing into a container field.
It seems to me that you are bright enough even before your morning coffee!!
The container field was completely overhauled in 12, but the import process was not. I know there is a lot of requests from the developer community for improvements to the import engine. I'd imagine future improvements could be possible that would let you filter your imports.