Github user kiszk commented on the issue: https://github.com/apache/spark/pull/19601 After I think about the choice for a while, I conclude that it is better to add the new `WritableColumnVector` (i.e. `UnsafeColumnVector`) and to keep the current `ColumnVector.Array`. I think that to add a new class will give us some flexibility and good abstraction between public class `ColumnVector` and other internal classes.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org