window functions with NULL literals in `partition by` and `order by` do not work: Internal("Sort operation is not applicable to scalar value NULL")

[alamb]

Reproducer:

echo "1" > /tmp/foo.csv
cargo run -p datafusion-cli

CREATE EXTERNAL TABLE foo(x int)
STORED AS CSV
LOCATION '/tmp/foo.csv';

> select percent_rank() OVER (partition by x) from foo;
+----------------+
| PERCENT_RANK() |
+----------------+
| 0              |
+----------------+
1 row in set. Query took 0.013 seconds.
> select percent_rank() OVER (partition by null) from foo;
ArrowError(ExternalError(ArrowError(ExternalError(Internal("Sort operation is not applicable to scalar value NULL")))))


> select percent_rank() OVER (partition by x order by null) from foo;
ArrowError(ExternalError(ArrowError(ExternalError(Internal("Sort operation is not applicable to scalar value NULL")))))

The correct output is null

See #1184 for more details

[alamb]

#1195 may be the same underlying cause

[xudong963]

#1195 may be the same underlying cause

@alamb Nice! It's the same cause.

> CREATE EXTERNAL TABLE foo(x int)
STORED AS CSV
LOCATION '/tmp/foo.csv';
0 rows in set. Query took 0.002 seconds.
> select percent_rank() OVER (partition by x) from foo;
+----------------+
| PERCENT_RANK() |
+----------------+
| 0              |
+----------------+
1 row in set. Query took 0.019 seconds.
> select percent_rank() OVER (partition by null) from foo;
+----------------+
| PERCENT_RANK() |
+----------------+
| 0              |
+----------------+
1 row in set. Query took 0.014 seconds.
> select percent_rank() OVER (partition by x order by null) from foo;
+----------------+
| PERCENT_RANK() |
+----------------+
| 0              |
+----------------+
1 row in set. Query took 0.014 seconds.
>

[xudong963]

Maybe the issue can close too.

[alamb]

As @xudong963 mentions, closed in #1197