Merge pull request #41 from awslabs/improve-exception-info

Improving exception info

Author: igorborgest

awswrangler/glue.py

@@ -4,7 +4,7 @@
 
 from awswrangler import data_types
 from awswrangler.athena import Athena
-from awswrangler.exceptions import UnsupportedFileFormat, InvalidSerDe, ApiError
+from awswrangler.exceptions import UnsupportedFileFormat, InvalidSerDe, ApiError, UnsupportedType
 
 logger = logging.getLogger(__name__)
 
@@ -192,7 +192,10 @@ def _build_schema(dataframe,
                 else:
                     schema_built.append((name, cast_columns[name]))
             else:
-                athena_type = data_types.pyarrow2athena(dtype)
+                try:
+                    athena_type = data_types.pyarrow2athena(dtype)
+                except UnsupportedType:
+                    raise UnsupportedType(f"Unsupported Pyarrow type for column {name}: {dtype}")
                 if name in partition_cols:
                     partition_cols_types[name] = athena_type
                 else:

testing/test_awswrangler/test_pandas.py

@@ -9,7 +9,7 @@
 import numpy
 
 from awswrangler import Session, Pandas
-from awswrangler.exceptions import LineTerminatorNotFound, EmptyDataframe, InvalidSerDe
+from awswrangler.exceptions import LineTerminatorNotFound, EmptyDataframe, InvalidSerDe, UnsupportedType
 
 logging.basicConfig(
     level=logging.INFO,
@@ -864,3 +864,21 @@ def test_to_parquet_duplicated_columns(
     assert len(list(dataframe.columns)) == len(list(dataframe2.columns))
     assert dataframe2.columns[0] == "a"
     assert dataframe2.columns[1] == "c"
+
+
+def test_to_parquet_with_pyarrow_null_type(
+        session,
+        bucket,
+        database,
+):
+    dataframe = pandas.DataFrame({
+        "a": [1, 2, 3],
+        "b": [4, 5, 6],
+        "col_null": [None, None, None],
+        "c": [7, 8, 9],
+    })
+    with pytest.raises(UnsupportedType):
+        assert session.pandas.to_parquet(dataframe=dataframe,
+                                  database=database,
+                                  path=f"s3://{bucket}/test/",
+                                  mode="overwrite")