Incorrect parsing of CSV file. Locale is ignored for some columns about dataframe HOT 2 OPEN

Ok, this is interesting. Here's the culprit:

val nf = NumberFormat.getInstance(Locale("et", "EE"))
nf.parse("-0,08")

This code throws an exception. Because in Estonian locale a different minus sign is used.
So DataFrame tries to parse this column using locale specific rules and fails. Next step of parsing is locale agnostic and it creates wrong values.
We need to improve UX in this part. The fact that locale parameter defines not only , and . is surprising. Maybe it shouldn't.. Suggested workaround: use ParserOptions(Locale.FRANCE)

from dataframe.

Note for future: let's check how it's handled in R dplyr and maybe use it as a reference. I think they might be good at interpretation of all kinds of data

from dataframe.