array type return different result

### Backend

VL (Velox)

### Bug description

The result shows different when running with array type.

Below is the reproduce case:
drop table array_test;
create table array_test stored as parquet as
select 1 as id, array() as a union all
select 2 as id, array("2A\n") as a union all
select 3 as id, array("3A\n","3B\n") as a union all
select 4 as id, array("4A\n","4B\n","4C\n") as a union all
select 5 as id, array() as a;

select * from array_test where id =4;

### Spark version

Spark-3.2.x

### Spark configurations

_No response_

### System information

_No response_

### Relevant logs

_No response_

Start by running the reported CREATE TABLE and SELECT reproduction with Spark 3.2.x and the Velox backend, then compare the result for id=4 with the expected array values. The issue provides no output, logs, test, or source file, so capture the differing results and trace the array handling before defining the fix and regression coverage.

Written by the indexing model from the issue text.