[SPARK-47002][Python] Return better error message if UDTF 'analyze' method 'orderBy' field accidentally returns a list of strings

[dtenedor]

What changes were proposed in this pull request?

This PR updates the Python UDTF API to check and return a better error message if the analyze method returns an AnalyzeResult object with an orderBy field erroneously set to a list of strings, rather than OrderingColumn instances.

For example, this UDTF accidentally sets the orderBy field in this way:

from pyspark.sql.functions import AnalyzeResult, OrderingColumn, PartitioningColumn
from pyspark.sql.types import IntegerType, Row, StructType
class Udtf:
    def __init__(self):
        self._partition_col = None
        self._count = 0
        self._sum = 0
        self._last = None

    @staticmethod
    def analyze(row: Row):
        return AnalyzeResult(
            schema=StructType()
                .add("user_id", IntegerType())
                .add("count", IntegerType())
                .add("total", IntegerType())
                .add("last", IntegerType()),
            partitionBy=[
                PartitioningColumn("user_id")
            ],
            orderBy=[
                "timestamp"
            ],
            )

    def eval(self, row: Row):
        self._partition_col = row["partition_col"]
        self._count += 1
        self._last = row["input"]
        self._sum += row["input"]

    def terminate(self):
        yield self._partition_col, self._count, self._sum, self._last

Why are the changes needed?

This improves error messages and helps keep users from getting confused.

Does this PR introduce any user-facing change?

Yes, see above.

How was this patch tested?

This PR adds test coverage.

Was this patch authored or co-authored using generative AI tooling?

No.

[dtenedor]

cc @ueshin here is the follow-up PR to add more checks for analyze result ordering fields.

[dtenedor]

The CI looks passing, the one failure is unrelated:

[ueshin]

Thanks! merging to master.