[HIVE-14454] Vectorization: Varchar() -> String should not copy underlying buffers

• Type: Improvement
• Status: Patch Available
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.2.0
• Component/s: Vectorization
• Assignee: Gopal Vijayaraghavan
• Reporter: Gopal Vijayaraghavan
• Created: 06/Aug/16 01:35
• Updated: 08/Aug/16 18:23

Description

The base byte[] array for a ByteColumnVector is immutable - all changes are made to a scratch column.

  @Override
  protected void func(BytesColumnVector outV, byte[][] vector, int[] start, int[] length, int i) {
    outV.setVal(i, vector[i], start[i], length[i]);
  }

the setVal copies the byte[] array, losing dictionary encoding in the process.