make description non Optional, correct docstring, optimize col conversion

Signed-off-by: varun-edachali-dbx <varun.edachali@databricks.com>

Author: varun-edachali-dbx

src/databricks/sql/result_set.py

@@ -512,17 +512,14 @@ def _convert_json_to_arrow(self, rows: List[List]) -> "pyarrow.Table":
         if not rows:
             return pyarrow.Table.from_pydict({})
 
-        columns = []
-        num_cols = len(rows[0])
-        for i in range(num_cols):
-            columns.append([row[i] for row in rows])
+        # Transpose rows to columns efficiently using zip
+        columns = list(map(list, zip(*rows)))
         names = [col[0] for col in self.description]
         return pyarrow.Table.from_arrays(columns, names=names)
 
     def _convert_json_types(self, rows: List[List]) -> List[List]:
         """
-        Convert raw data rows to Row objects with named columns based on description.
-        Also converts string values to appropriate Python types based on column metadata.
+        Convert string values to appropriate Python types based on column metadata.
         """
 
         if not self.description or not rows:

src/databricks/sql/utils.py

@@ -274,7 +274,7 @@ def __init__(
         start_row_offset: int = 0,
         result_links: Optional[List[TSparkArrowResultLink]] = None,
         lz4_compressed: bool = True,
-        description: Optional[List[Tuple]] = None,
+        description: List[Tuple] = [],
     ):
         """
         A queue-like wrapper over CloudFetch arrow batches.