test: Add data file for JSON dtype system tests (#1136)

Author: chelsea-lin

tests/data/json.jsonl

@@ -0,0 +1,16 @@
+{"rowindex": 0, "json_col": null}
+{"rowindex": 1, "json_col": true}
+{"rowindex": 2, "json_col": 100}
+{"rowindex": 3, "json_col": 0.98}
+{"rowindex": 4, "json_col": "a string"}
+{"rowindex": 5, "json_col": []}
+{"rowindex": 6, "json_col": [1, 2, 3]}
+{"rowindex": 7, "json_col": [{"a": 1}, {"a": 2}, {"a": null}, {}]}
+{"rowindex": 8, "json_col": {"bool_value": true}}
+{"rowindex": 9, "json_col": {"folat_num": 3.14159}}
+{"rowindex": 10, "json_col": {"date": "2024-07-16"}}
+{"rowindex": 11, "json_col": {"null_filed": null}}
+{"rowindex": 12, "json_col": {"int_value": 2, "null_filed": null}}
+{"rowindex": 13, "json_col": {"list_data": [10, 20, 30]}}
+{"rowindex": 14, "json_col": {"person": {"name": "Alice", "age": 35}}}
+{"rowindex": 15, "json_col": {"order": {"items": ["book", "pen"], "total": 15.99}}}

tests/data/json_schema.json

@@ -0,0 +1,12 @@
+[
+    {
+        "name": "rowindex",
+        "type": "INTEGER",
+        "mode": "REQUIRED"
+    },
+    {
+        "name": "json_col",
+        "type": "JSON",
+        "mode": "NULLABLE"
+    }
+]

tests/system/conftest.py

@@ -298,6 +298,7 @@ def load_test_data_tables(
         ("nested", "nested_schema.json", "nested.jsonl"),
         ("nested_structs", "nested_structs_schema.json", "nested_structs.jsonl"),
         ("repeated", "repeated_schema.json", "repeated.jsonl"),
+        ("json", "json_schema.json", "json.jsonl"),
         ("penguins", "penguins_schema.json", "penguins.jsonl"),
         ("time_series", "time_series_schema.json", "time_series.jsonl"),
         ("hockey_players", "hockey_players.json", "hockey_players.jsonl"),
@@ -384,6 +385,11 @@ def repeated_table_id(test_data_tables) -> str:
     return test_data_tables["repeated"]
 
 
+@pytest.fixture(scope="session")
+def json_table_id(test_data_tables) -> str:
+    return test_data_tables["json"]
+
+
 @pytest.fixture(scope="session")
 def penguins_table_id(test_data_tables) -> str:
     return test_data_tables["penguins"]
@@ -481,6 +487,25 @@ def repeated_pandas_df() -> pd.DataFrame:
     return df
 
 
+@pytest.fixture(scope="session")
+def json_df(
+    json_table_id: str, session: bigframes.Session
+) -> bigframes.dataframe.DataFrame:
+    """Returns a DataFrame containing columns of JSON type."""
+    return session.read_gbq(json_table_id, index_col="rowindex")
+
+
+@pytest.fixture(scope="session")
+def json_pandas_df() -> pd.DataFrame:
+    """Returns a DataFrame containing columns of JSON type."""
+    df = pd.read_json(
+        DATA_DIR / "json.jsonl",
+        lines=True,
+    )
+    df = df.set_index("rowindex")
+    return df
+
+
 @pytest.fixture(scope="session")
 def scalars_df_default_index(
     scalars_df_index: bigframes.dataframe.DataFrame,

tests/system/small/test_series.py

@@ -278,6 +278,13 @@ def test_get_column(scalars_dfs, col_name, expected_dtype):
     assert series_pandas.shape[0] == scalars_pandas_df.shape[0]
 
 
+def test_get_column_w_json(json_df, json_pandas_df):
+    series = json_df["json_col"]
+    series_pandas = series.to_pandas()
+    assert series.dtype == pd.StringDtype(storage="pyarrow")
+    assert series_pandas.shape[0] == json_pandas_df.shape[0]
+
+
 def test_series_get_column_default(scalars_dfs):
     scalars_df, _ = scalars_dfs
     result = scalars_df.get(123123123123123, "default_val")