Hm, this is the query I used to export the data, I don't really see any way the data could have been incomplete from this. Maybe @gabriel-wmde / @Tonina_Zhelyazkova_WMDE can have another look at this and see if they notice anything odd:
hive -e 'SELECT year, month, day, event.bannerName, event.finalSlide, event.slidesShown, useragent.browser_family, useragent.browser_major, useragent.device_family, useragent.os_family, useragent.os_major, wiki FROM event.wmdebannerevents WHERE year = 2018 AND event.bannerAction = "banner-closed";' | sed 's/[\t]/,/g' > ~/banner_close_events_2018.csv
I have done another export and sent another email with the latest dump. It has a bit different file size but it may be that I cut off a tiny bit myself last time. Please check if that resolves the issue.
Other than that, my best guess would be that maybe the data was too much and the software you used (Microsoft Office?) to work with it cut parts of it off. I cannot open the complete file in LibreOffice, for example.