Lost transcription when importing from .csv

To whom may have a moment to answer,

I’m using Elan 6.2 and I’ve never seen the following kind of behaviour before.

For my study I need to export Latin text from Elan, transliterate it into Cyrillic using a third-party software, and then import it back into Elan. I am trying to import Cyrillic from a .csv file.

First weird observation is that when I import it, it separates text after a comma into a subsequent tier. In the instances where text is missing, Elan will simply ‘lose’ text after a second and/or third comma from an individual line in .csv.

Please advise how to trouble shoot this


Yes, although importing a .csv file is supported now, this import has its limitations. It is still more or less the tab-delimited text import function, accepting .csv files as input and using the , as the separator. The importer does not detect pieces of text in double quotes (") first, before splitting a line into columns. Therefore a comma in an annotation is interpreted as a delimiter, messing up the alignment. Apart from that
double quotes are not removed and escaped double quotes within annotations are also not detected.

It is probably best to save the file as a tab-delimited text .txt file and import that into ELAN.
I’m not sure if this also solves the second problem; in principle missing text should be dealt with correctly if the delimiters are there.