Implement parquet_metadata function in datafusion-cli

[alamb]

Is your feature request related to a problem or challenge?

When exploring Parquet files using datafusion-cli I would often like to see how they are structured (how many row groups, if they have statistics, etc).

Describe the solution you'd like

I would like to create new functions for exploring parquet metadata using the new User Defined Table Functions (🙇 to @Veeupup ) introduced in #8306

Ideally we could implement something like parquet_metadata: https://duckdb.org/docs/data/parquet/overview

(I think parquet_schema is covered by describe 'filename.parquet' already

Note I think this should be done in datafusion-cli (not core DataFusion)

Describe alternatives you've considered

No response

Additional context

No response

[alamb]

This would also be a great test of the user defined table function feature to see if we can build something slightly more complicated than read_csv

[Veeupup]

I can help with this ticket as a following PR to #8306

[Veeupup]

After this ticker was finished, maybe we can have a list of internal table functions to implement, just like:

• read_csv
• read_parquet
• read_json
• ...

[alamb]

I am not sure about read_csv etc -- the datafusion-cli already has the ability to do the equivalent of read_csv (and the other formats) with a different format

Instead of

select * from read_csv('foo.csv')

It uses

select * from 'foo.csv'