ARROW-11594: [Rust] Support pretty printing of NullArray

[alamb]

As pointed out by @tustvold on https://github.com/influxdata/influxdb_iox/pull/783#discussion_r574414243, it would seem the whole point of NullArray::new_with_type, is to cheaply construct entirely null columns, with a smaller memory footprint.

Currently trying to print them out causes a panic:

    #[test]
    fn test_pretty_format_null() -> Result<()> {
        // define a schema.
        let schema = Arc::new(Schema::new(vec![
            Field::new("a", DataType::Utf8, true),
            Field::new("b", DataType::Int32, true),
        ]));

        let num_rows = 4;

        // define data (null)
        let batch = RecordBatch::try_new(
            schema,
            vec![
                Arc::new(NullArray::new_with_type(num_rows, DataType::Utf8)),
                Arc::new(NullArray::new_with_type(num_rows, DataType::Int32)),
            ],
        )?;

        let table = pretty_format_batches(&[batch])?;
}

Panics:

failures:

---- util::pretty::tests::test_pretty_format_null stdout ----
thread 'util::pretty::tests::test_pretty_format_null' panicked at 'called `Option::unwrap()` on a `None` value', arrow/src/util/display.rs:201:27

Note

_Update: the issue with NullArray claiming to be types such as Int32 has been fixed by @nevi-me in #9469 _

[github-actions]

https://issues.apache.org/jira/browse/ARROW-11594

[nevi-me]

heads-up @alamb, the implementation of NullArray::new_with_type is not sound, because it's not compliant with the spec.
A NullArray must always have DataType::Null.

I was working on a solution for this last night, but didn't finish it; I'll wrap it up tonight. This mainly fixes #9331. [EDIT: I've opened https://github.com//pull/9469].

This PR will still be necessary, as we'll need to be able to print NullArray correctly (but haven't looked at the code)

[paddyhoran]

Sorry, I didn't scroll and see the comment from @nevi-me. It only affects your test but I guess we should resolve these items first.

[alamb]

Thanks @paddyhoran -- I think the logic in this particular PR is sound (with the possibly exception that NullArray::new_with_type is going away) as it just checks for nulls and adds a test :) I am going over to lend a hand with #9469 and hopefully we'll have something good

[alamb]

Switching back to draft until #9469 gets merged

[alamb]

Rebased to pick up #9469

[jorgecarleitao]

lgtm. Thanks @alamb !

I agree with @nevi-me : Array::data_type(&a) == DataType::Null if and only if a.as_ref().downcast_ref() is infalible.

@alamb , the general solution here is to have a new entry in match data_type for DataType::Null. In general, every missing branch on these matches is either an unimplemented feature or unreachable. :)

[alamb]

I agree with @nevi-me : Array::data_type(&a) == DataType::Null if and only if a.as_ref().downcast_ref() is infalible.

Good call @jorgecarleitao -- I didn't realize that the real bug was NullArray created by NullArray::new_with_type claiming to be DataType::Int32, etc. Since this has been fixed by @nevi-me in #9469 I am feeling much better and have updated this PR's description to remove the "Note" section