As a few others have mentioned, the parsing alone means DOCX is a huge win over PDF. I had to parse a bunch of PDF data related to COVID and it was always a PITA. Every time they changed their layout even a little bit I had to rewrite parts of my extractor. The worst part? The headers/metadata showed it was all made in Word so they could have exported to DOCX as well as PDF if they wanted to but they only provided PDF.