ARROW-9680: [Java] Support non-nullable vectors

[liyafan82]

This issue was first discussed in the ML (https://lists.apache.org/thread.html/r480387ec9ec822f3ed30e9131109e43874a1c4d18af74ede1a7e41c5%40%3Cdev.arrow.apache.org%3E), from which we have received some feedback.

We briefly resate it here as below:

1. Non-nullable vectors are widely used in practice. For example, in a database engine, a column can be declared as not null, so it cannot contain null values.
   2.Non-nullable vectors has significant performance advantages compared with their nullable conterparts, such as:

1. the memory space of the validity buffer can be saved.
2. manipulation of the validity buffer can be bypassed
3. some if-else branches can be replaced by sequential instructions (by the JIT compiler), leading to high throughput for the CPU pipeline.

We open this Jira to facilitate further discussions, and we may provide a sample PR, which will help us make a clearer decision.

[liyafan82]

In the first commit of this PR, we only support IntVector.

[github-actions]

https://issues.apache.org/jira/browse/ARROW-9680

[kiszk]

Can we declare NON_NULLABLE_VECTORS_ENABLED && nullable as a final instance variable? If so, we can avoid a similar conditional expression.

[liyafan82]

@kiszk Thanks for your feedback. I have changed the code accordingly.
It really makes the code simpler, and I've verified that there were no performance regression.

However, it could also lead to some consufing behavior. For example, a vector that is known to have no null values may have nullable field equal to true.

[emkornfield]

I think consensus on the mailing list is that this isn't something we want to pursue?

[liyafan82]

I think consensus on the mailing list is that this isn't something we want to pursue?

@emkornfield Thanks for your kind reminder. I am closing it.