Varchar datatype is also saved internally as a String. Now the Hive Varchar does not support vectorization. Vectorization allows Hive to process a batch of rows together instead of processing one row at a time. Operations are performed on the entire column vector which improves the instruction pipelines and cache usage.