fix: handle unsupported types and empty results in describe (#2506)

The .describe() method attempts to compute metrics like mean, max, and
unique on all columns. For MM complex types (structs representing
images/videos), running COUNT(DISTINCT column) or mathematical
aggregates is physically prohibited by BigQuery and raises syntax/type
validation errors.
We limit aggregations on OBJ_REF_DTYPE and JSON columns to only a basic
.count(), skipping unhashable/unsupported summary metrics.
Fixes #<452681068> 🦕

Author: shuoweil

bigframes/pandas/core/methods/describe.py

@@ -56,9 +56,10 @@ def describe(
 "max",
  ]
  ).intersection(describe_block.column_labels.get_level_values(-1))
-describe_block=describe_block.stack(override_labels=stack_cols)
-
-returndataframe.DataFrame(describe_block).droplevel(level=0)
+ifnotstack_cols.empty:
+describe_block=describe_block.stack(override_labels=stack_cols)
+returndataframe.DataFrame(describe_block).droplevel(level=0)
+returndataframe.DataFrame(describe_block)
 
 
 def_describe(
@@ -120,5 +121,7 @@ def _get_aggs_for_dtype(dtype) -> list[aggregations.UnaryAggregateOp]:
 dtypes.TIME_DTYPE,
  ]:
 return [aggregations.count_op, aggregations.nunique_op]
+elifdtypes.is_json_like(dtype) ordtype==dtypes.OBJ_REF_DTYPE:
+return [aggregations.count_op]
 else:
 return []

tests/system/small/pandas/test_describe.py

@@ -15,6 +15,8 @@
 importpandas.testing
 importpytest
 
+importbigframes.pandasasbpd
+
 
 deftest_df_describe_non_temporal(scalars_dfs):
 # TODO: supply a reason why this isn't compatible with pandas 1.x
@@ -352,3 +354,40 @@ def test_series_groupby_describe(scalars_dfs):
 check_dtype=False,
 check_index_type=False,
  )
+
+
+deftest_describe_json_and_obj_ref_returns_count(session):
+# Test describe() works on JSON and OBJ_REF types (without nunique, which fails)
+sql="""
+ SELECT
+ PARSE_JSON('{"a": 1}') AS json_col,
+ 'gs://cloud-samples-data/vision/ocr/sign.jpg' AS uri_col
+ """
+df=session.read_gbq(sql)
+
+df["obj_ref_col"] =df["uri_col"].str.to_blob()
+df=df.drop(columns=["uri_col"])
+
+res=df.describe(include="all").to_pandas()
+
+assert"count"inres.index
+assertres.loc["count", "json_col"] ==1.0
+assertres.loc["count", "obj_ref_col"] ==1.0
+
+
+deftest_describe_with_unsupported_type_returns_empty_dataframe(session):
+df=session.read_gbq("SELECT ST_GEOGPOINT(1.0, 2.0) AS geo_col")
+
+res=df.describe().to_pandas()
+
+assertlen(res.columns) ==0
+assertlen(res.index) ==1
+
+
+deftest_describe_empty_dataframe_returns_empty_dataframe(session):
+df=bpd.DataFrame()
+
+res=df.describe().to_pandas()
+
+assertlen(res.columns) ==0
+assertlen(res.index) ==1