From 915884dda35bd11418aae6319b30b4cfdb449d2d Mon Sep 17 00:00:00 2001
From: Elizabeth Thompson <eschutho@gmail.com>
Date: Fri, 2 Dec 2022 13:40:20 -0800
Subject: [PATCH] convert <NA> values to None instead of stringifying

---
 superset/result_set.py              |  6 ++++-
 tests/unit_tests/result_set_test.py | 41 +++++++++++++++++++++++++++++
 2 files changed, 46 insertions(+), 1 deletion(-)

diff --git a/superset/result_set.py b/superset/result_set.py
index cf57ddc84af54..a6f39f76bf4df 100644
--- a/superset/result_set.py
+++ b/superset/result_set.py
@@ -67,7 +67,11 @@ def stringify_values(array: np.ndarray) -> np.ndarray:
 
     with np.nditer(result, flags=["refs_ok"], op_flags=["readwrite"]) as it:
         for obj in it:
-            obj[...] = stringify(obj)
+            if pd.isna(obj):
+                # pandas <NA> type cannot be converted to string
+                obj[pd.isna(obj)] = None
+            else:
+                obj[...] = stringify(obj)
 
     return result
 
diff --git a/tests/unit_tests/result_set_test.py b/tests/unit_tests/result_set_test.py
index 48b9576a4ca79..e7371f5c0fa3d 100644
--- a/tests/unit_tests/result_set_test.py
+++ b/tests/unit_tests/result_set_test.py
@@ -18,6 +18,13 @@
 # pylint: disable=import-outside-toplevel, unused-argument
 
 
+import numpy as np
+import pandas as pd
+from numpy.core.multiarray import array
+
+from superset.result_set import stringify_values
+
+
 def test_column_names_as_bytes() -> None:
     """
     Test that we can handle column names as bytes.
@@ -65,3 +72,37 @@ def test_column_names_as_bytes() -> None:
 |  1 | 2016-01-27 | 392.444 | 396.843 | 391.782 | 394.972 |     394.972 | 47424400 |
     """.strip()
     )
+
+
+def test_stringify_with_null_integers():
+    """
+    Test that we can safely handle type errors when an integer column has a null value
+    """
+
+    data = [
+        ("foo", "bar", pd.NA, None),
+        ("foo", "bar", pd.NA, True),
+        ("foo", "bar", pd.NA, None),
+    ]
+    numpy_dtype = [
+        ("id", "object"),
+        ("value", "object"),
+        ("num", "object"),
+        ("bool", "object"),
+    ]
+
+    array2 = np.array(data, dtype=numpy_dtype)
+    column_names = ["id", "value", "num", "bool"]
+
+    result_set = np.array([stringify_values(array2[column]) for column in column_names])
+
+    expected = np.array(
+        [
+            array(['"foo"', '"foo"', '"foo"'], dtype=object),
+            array(['"bar"', '"bar"', '"bar"'], dtype=object),
+            array([None, None, None], dtype=object),
+            array([None, "true", None], dtype=object),
+        ]
+    )
+
+    assert np.array_equal(result_set, expected)