Just curious — was there a specific reason you chose not to use the encoding/csv package?
Using it might simplify the parsing logic and make the code a bit more robust and readable. Unless of course your approach was to implement your own parser for a learning opportunity then that is fine too