We recently did a batch upload of a whole run of a journal and noticed during proofing that a number of article titles had gone missing in the upload. Our import XML file contained text, and our original XML contained text, in all the appropriate places. On the ToC, the author name was visible, the abstract and link to HTML were visible, but the title was not.
On further inspection we noticed that all the titles had disappeared from articles with titles containing two single quotation marks. If we deleted the single quotes (or only removed one of the two single quotes) from the import XML file then the title would appear, but obviously, we'd like to not have to double back just to fix metadata that the import file should handle.
Jason Nugent and I talked about it for a while and we think that maybe somewhere along the line, the import program isn't properly escaping single quotation mark characters. If there are two of them in a title, it will disappear.
Also, I'd point out, it's really hard to fix this within OJS, because you can't get to the file summary without being able to click on the title of the article, and if the article has no title, there's no way to edit the metadata within OJS.