Subject: | Text::CSV_PP and blank_is_undef: problem if File is flagged as being UTF-8-encoded |
Quote from http://stackoverflow.com/a/33935166/198183:
"This appears to be a bug in Text::CSV_PP (Text::CSV_XS doesn't display this symptom)
The problem is that, if the line read from the file is flagged as being UTF-8-encoded, then each field split using the separator character is passed through utf8::encode even if it has previously been set to undef because the field is blank and the blank_is_undef option is enabled. That incorrectly sets it back to an empty string"