Fix toUpperCase/toLowerCase to use Locale.ENGLISH

String#toLowerCase()/toUpperCase() without a locale argument can occur unexpected behavior based on the locale. It's written in Javadoc:

For instance, "TITLE".toLowerCase() in a Turkish locale returns "t\u0131tle", where '\u0131' is the LATIN SMALL LETTER DOTLESS I character

HADOOP-11602 is an issue to address this problem at the Hadoop side.