The file produced by “Export As > Tab-delimited text” has a spurious tab after the name of the last column. (See the screenshot: the right-arrow is how Notepad++ represented the tab character.)
This is a bit painful when importing the file into some other software. For example when loading it into R using ‘readr’, you get an extra anonymous column; when loading it into LaTeX using ‘csvsimple’, you need to either ignore the header or suppress column count checking.
Ideally, the last column name in the header should be followed immediately by an end-of-line, with no tab: just like in the individual non-header rows of the exported transcript, there is no tab after the last column data.
I’m using ELAN 6.3 on Windows 10.