Skip to content

A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.

fix: table filtering by NaN in pandas string columns (#9092)

Kiran Gadhave55b1f86b2ef6
diff --git a/marimo/_plugins/ui/_impl/tables/narwhals_table.py b/marimo/_plugins/ui/_impl/tables/narwhals_table.py
index 4f99a6c4b3..e05d48b057 100644
--- a/marimo/_plugins/ui/_impl/tables/narwhals_table.py
+++ b/marimo/_plugins/ui/_impl/tables/narwhals_table.py
@@ -237,14 +237,19 @@ class NarwhalsTableManager(
result = _calculate_top_k_rows(frame)
value_counts: list[tuple[Any, int]] = []
-
- # NaNs and Infs serialize to null, which isn't distingushable from normal nulls
- # so instead we set to string values
+ col_dtype = self.data.collect_schema()[column]
+
+ # NaNs and Infs serialize to null, which isn't distinguishable from
+ # normal nulls, so we replace them with string tokens — but only for
+ # float columns where NaN/Inf are meaningful distinct values.
+ # For non-float columns (e.g. strings), NaN is just pandas' missing
+ # value sentinel and should stay as None.
+ is_float_col = col_dtype.is_float()
for row in result.rows():
value = unwrap_py_scalar(row[0])
count = int(unwrap_py_scalar(row[1]))
if isinstance(value, float) and math.isnan(value):
- value = NAN_VALUE
+ value = NAN_VALUE if is_float_col else None
elif isinstance(value, float) and math.isinf(value) and value > 0:
value = POSITIVE_INF
elif isinstance(value, float) and math.isinf(value) and value < 0:
@@ -278,9 +283,7 @@ class NarwhalsTableManager(
return ("time", dtype_string)
elif dtype == nw.Date:
return ("date", dtype_string)
- elif dtype == nw.Datetime:
- return ("datetime", dtype_string)
- elif dtype.is_temporal():
+ elif dtype == nw.Datetime or dtype.is_temporal():
return ("datetime", dtype_string)
elif dtype.is_numeric():
return ("number", dtype_string)
@@ -463,35 +466,7 @@ class NarwhalsTableManager(
),
}
)
- elif is_narwhals_integer_type(dtype):
- exprs.update(
- {
- "unique": col.n_unique(),
- "min": col.min(),
- "max": col.max(),
- "mean": col.mean(),
- "std": col.std(),
- "median": col.median(),
- }
- )
- if supports_numeric_quantiles:
- exprs.update(
- {
- "p5": col.quantile(
- 0.05, interpolation=quantile_interpolation
- ),
- "p25": col.quantile(
- 0.25, interpolation=quantile_interpolation
- ),
- "p75": col.quantile(
- 0.75, interpolation=quantile_interpolation
- ),
- "p95": col.quantile(
- 0.95, interpolation=quantile_interpolation
- ),
- }
- )
- elif dtype.is_numeric():
+ elif is_narwhals_integer_type(dtype) or dtype.is_numeric():
exprs.update(
{
"unique": col.n_unique(),
diff --git a/tests/_plugins/ui/_impl/dataframes/test_handlers.py b/tests/_plugins/ui/_impl/dataframes/test_handlers.py
index 0312c6152b..95e2afca2e 100644
--- a/tests/_plugins/ui/_impl/dataframes/test_handlers.py
+++ b/tests/_plugins/ui/_impl/dataframes/test_handlers.py
@@ -1921,9 +1921,6 @@ class TestTransformHandler:
assert_frame_equal_with_nans(result, expected)
@staticmethod
- @pytest.mark.xfail(
- reason="NaN filtering for object dtypes in pandas aren't implemented"
- )
@pytest.mark.parametrize(
("df", "expected"),
list(
@@ -1931,9 +1928,7 @@ class TestTransformHandler:
create_test_dataframes(
{"nulls": [1, 2, 3, None, "hello"]}, include=["pandas"]
),
- create_test_dataframes(
- {"nulls": [float("nan")]}, include=["pandas"]
- ),
+ create_test_dataframes({"nulls": [None]}, include=["pandas"]),
)
),
)
@@ -1947,12 +1942,12 @@ class TestTransformHandler:
Condition(
column_id="nulls",
operator="in",
- value=[NAN_VALUE],
+ value=[None],
)
],
)
result = apply(df, in_transform)
- assert_frame_equal(result, expected)
+ assert_frame_equal_with_nans(result, expected)
@staticmethod
@pytest.mark.parametrize(