Create DDL Command `DESCRIBE/EXPLAIN/SHOW`

[comphead]

The idea is to output the schema for query similar to DuckDB or Apache Spark

D describe select map_extract(map([1], ['a']), 1)[1];
┌────────────────────────────────────────────────────────────────────┬─────────────┬─────────┬─────────┬─────────┬─────────┐
│                            column_name                             │ column_type │  null   │   key   │ default │  extra  │
│                              varchar                               │   varchar   │ varchar │ varchar │ varchar │ varchar │
├────────────────────────────────────────────────────────────────────┼─────────────┼─────────┼─────────┼─────────┼─────────┤
│ map_extract("map"(main.list_value(1), main.list_value('a')), 1)[1] │ VARCHAR     │ YES     │         │         │         │
└────────────────────────────────────────────────────────────────────┴─────────────┴─────────┴─────────┴─────────┴─────────┘
D

it would require some discussion what command to use? Some possible options below

Possible DDL options

DESCRIBE query - like in DuckDB
EXPLAIN SCHEMA query
SHOW SCHEMA query

[alamb]

We already use describe foo for table schema

DataFusion CLI v49.0.2
> create table foo (x int) as values (1);
0 row(s) fetched.
Elapsed 0.033 seconds.

> describe foo;
+-------------+-----------+-------------+
| column_name | data_type | is_nullable |
+-------------+-----------+-------------+
| x           | Int32     | YES         |
+-------------+-----------+-------------+
1 row(s) fetched.
Elapsed 0.006 seconds.

Describe appears to be an alias of explain:

> describe  select x+1 from foo;
+---------------+-------------------------------+
| plan_type     | plan                          |
+---------------+-------------------------------+
| physical_plan | ┌───────────────────────────┐ |
|               | │       ProjectionExec      │ |
|               | │    --------------------   │ |
|               | │     foo.x + Int64(1):     │ |
|               | │    CAST(x AS Int64) + 1   │ |
|               | └─────────────┬─────────────┘ |
|               | ┌─────────────┴─────────────┐ |
|               | │       DataSourceExec      │ |
|               | │    --------------------   │ |
|               | │         bytes: 112        │ |
|               | │       format: memory      │ |
|               | │          rows: 1          │ |
|               | └───────────────────────────┘ |
|               |                               |
+---------------+-------------------------------+
1 row(s) fetched.
Elapsed 0.001 seconds.

Maybe we can follow the model of whatever system that comes from 🤔

[comphead]

Maybe we can follow the model of whatever system that comes from 🤔

The closest is DESCRIBE query like DuckDB does,

Apache Spark uses a dataframe API method like

spark.sql("select 1").printSchema()

which doesn't seem usable in SQL context which more fits to DataFusion. Other major query engines doesn't support this functionality. Therefore I'll probably start with DuckDb's DESCRIBE query if no other suggestions

[comphead]

cc @djanderson as working on #18234

The idea is to output Spark like tree view schema representation