Change default value of datafusion.catalog.has_header to true

[alamb]

Is your feature request related to a problem or challenge?

As @jgranduel notes in #11848

I think most people would expect when reading this data as a csv file, we would get two columns a and b

a,b
1,2

However, datafusion currently treats the CSV file as "without header" by default resulting in

> datafusion-cli.exe
DataFusion CLI v40.0.0
> select * from './data_dir/';
+----------+----------+
| column_1 | column_2 |
+----------+----------+
| a        | b        |
| 1        | 2        |
+----------+----------+
4 row(s) fetched.

Describe the solution you'd like

I propose removing this "rough edge" and changing the datafusion default behavior to be what I think most people would expect

+---+---+
| a | b |
+---+---+
| 1 | 2 |
+---+---+

Describe alternatives you've considered

I think we should change the default vale of datafusion.catalog.has_header from false to true

datafusion/datafusion/common/src/config.rs

Lines 184 to 186 in 79fa6f9

	/// Default value for `format.has_header` for `CREATE EXTERNAL TABLE`
	/// if not specified explicitly in the statement.
	pub has_header: bool, default = false

Additional context

No response

[korowa]

I think it'll be fine because there is already CSVReadOptions with default has_header: true (link), which produces some inconsistency in default handling csv.

[alamb]

I think default to "no headers" is unlikely to have been a conscious choice (I suspect it was a historical accident)