Support writing a table with partition column(s) of type LargeUtf8

[paleolimbot]

Is your feature request related to a problem or challenge?

The freshly released Pandas 3.0 has a new string dtype that defaults to LargeUtf8 when converted to Arrow. There may be a few things this triggers with respect to LargeUtf8 view support; however the one that caused a failing test for us was support for LargeUtf8 when writing Parquet with partitions ( apache/sedona-db#538 ). The error is it is not yet supported to write to hive partitions with datatype LargeUtf8.

Describe the solution you'd like

I think it would be fairly easy to add a branch here to support it. I'm happy to do this.

datafusion/datafusion/datasource/src/write/demux.rs

Lines 394 to 405 in 9f27e93

	DataType::Utf8 => {
	let array = as_string_array(col_array)?;
	for i in 0..rb.num_rows() {
	partition_values.push(Cow::from(array.value(i)));
	}
	}
	DataType::Utf8View => {
	let array = as_string_view_array(col_array)?;
	for i in 0..rb.num_rows() {
	partition_values.push(Cow::from(array.value(i)));
	}
	}

Describe alternatives you've considered

Pandas should also probably consider just sticking with Utf8 as the default conversion to arrow (there are probably also other places/libraries that haven't supported this all the way yet).

Additional context

No response