When the Integration Service uses the Unicode data movement mode, the Integration Service reads the precision of Char, Varchar, and Clob columns based on the length semantics you set for columns in the Oracle database. If you use the byte semantics to determine column length, the Integration Service reads the precision as the number of bytes. If you use the char semantics, the Integration Service reads the precision as the number of characters.