fix a PX input bug when using data frame indices

Author: nicolaskruchten

packages/python/plotly/plotly/express/_core.py

@@ -954,7 +954,7 @@ def build_dataframe(args, attrables, array_attrables):
                         )
                     )
                 col_name = str(argument)
-                df_output[col_name] = df_input[argument]
+                df_output[col_name] = df_input[argument].values
             # ----------------- argument is a column / array / list.... -------
             else:
                 is_index = isinstance(argument, pd.RangeIndex)
@@ -989,7 +989,10 @@ def build_dataframe(args, attrables, array_attrables):
                         "length of previous arguments %s is %d"
                         % (field, len(argument), str(list(df_output.columns)), length)
                     )
-                df_output[str(col_name)] = argument
+                if hasattr(argument, "values"):
+                    df_output[str(col_name)] = argument.values
+                else:
+                    df_output[str(col_name)] = np.array(argument)
 
             # Finally, update argument with column name now that column exists
             if field_name not in array_attrables:

packages/python/plotly/plotly/tests/test_core/test_px/test_px_input.py

@@ -2,8 +2,6 @@
 import numpy as np
 import pandas as pd
 import pytest
-import plotly.graph_objects as go
-import plotly
 from plotly.express._core import build_dataframe
 from pandas.util.testing import assert_frame_equal
 
@@ -234,6 +232,24 @@ def test_build_df_with_index():
     assert_frame_equal(tips.reset_index()[out["data_frame"].columns], out["data_frame"])
 
 
+def test_non_matching_index():
+    df = pd.DataFrame(dict(y=[1, 2, 3]), index=["a", "b", "c"])
+
+    expected = pd.DataFrame(dict(x=["a", "b", "c"], y=[1, 2, 3]))
+
+    args = dict(data_frame=df, x=df.index, y="y")
+    out = build_dataframe(args, all_attrables, array_attrables)
+    assert_frame_equal(expected, out["data_frame"])
+
+    args = dict(data_frame=None, x=df.index, y=df.y)
+    out = build_dataframe(args, all_attrables, array_attrables)
+    assert_frame_equal(expected, out["data_frame"])
+
+    args = dict(data_frame=None, x=["a", "b", "c"], y=df.y)
+    out = build_dataframe(args, all_attrables, array_attrables)
+    assert_frame_equal(expected, out["data_frame"])
+
+
 def test_splom_case():
     iris = px.data.iris()
     fig = px.scatter_matrix(iris)