With pandas to read an Oracle database, content is in Chinese, no problem when reading a few columns (including Chinese), read a column will appear unicodeerror, how to deal with this?