accept duplicate column names in table report

CHANGES.rst

@@ -72,6 +72,10 @@ Bug fixes
   fail on polars dataframes when used with the default parameters. This has been
   fixed in :pr:`1122` by :user:`Jérôme Dockès <jeromedockes>`.
 
+* The :class:`TableReport` would raise an exception when the input (pandas)
+  dataframe contained several columns with the same name. This has been fixed in
+  :pr:`1125` by :user:`Jérôme Dockès <jeromedockes>`.
+
 Release 0.3.1
 =============
 

skrub/_column_associations.py

@@ -104,8 +104,8 @@ def _onehot_encode(df, n_bins):
     """
     n_rows, n_cols = sbd.shape(df)
     output = np.zeros((n_cols, n_bins, n_rows), dtype=bool)
-    for col_idx, col_name in enumerate(sbd.column_names(df)):
-        col = sbd.col(df, col_name)
+    for col_idx in range(n_cols):
+        col = sbd.col_by_idx(df, col_idx)
         if sbd.is_numeric(col):
             col = sbd.to_float32(col)
             if _CATEGORICAL_THRESHOLD <= sbd.n_unique(col):

skrub/_dataframe/_common.py

@@ -40,6 +40,7 @@
     "is_column_list",
     "to_column_list",
     "col",
+    "col_by_idx",
     "collect",
     #
     # Querying and modifying metadata
@@ -350,7 +351,7 @@
     if is_column(obj):
         return [obj]
     if is_dataframe(obj):
-        return [col(obj, c) for c in column_names(obj)]
+        return [col_by_idx(obj, idx) for idx in range(shape(obj)[1])]
     if not is_column_list(obj):
         raise TypeError("obj should be a DataFrame, a Column or a list of Columns.")
     return obj
@@ -371,6 +372,21 @@
     return df[col_name]
 
 
+@dispatch
+def col_by_idx(df, col_idx):
+    raise NotImplementedError()
+
+
+@col_by_idx.specialize("pandas", argument_type="DataFrame")
+def _col_by_idx_pandas(df, col_idx):
+    return df.iloc[:, col_idx]
+
+
+@col_by_idx.specialize("polars", argument_type="DataFrame")
+def _col_by_idx_polars(df, col_idx):
+    return df[df.columns[col_idx]]
+
+
 @dispatch
 def collect(df):
     return df

skrub/_dataframe/tests/test_common.py

@@ -228,6 +228,16 @@ def test_to_column_list(df_module, example_data_dict):
         ns.to_column_list(None)
 
 
+def test_to_column_list_duplicate_columns(pd_module):
+    df = pd_module.make_dataframe({"a": [1, 2], "b": [3, 4]})
+    df.columns = ["a", "a"]
+    col_list = ns.to_column_list(df)
+    assert ns.name(col_list[0]) == "a"
+    assert ns.to_list(col_list[0]) == [1, 2]
+    assert ns.name(col_list[1]) == "a"
+    assert ns.to_list(col_list[1]) == [3, 4]
+
+
 def test_collect(df_module):
     assert ns.collect(df_module.example_dataframe) is df_module.example_dataframe
     if df_module.name == "polars":
@@ -236,6 +246,20 @@ def test_collect(df_module):
         )
 
 
+def test_col(df_module):
+    assert ns.to_list(ns.col(df_module.example_dataframe, "float-col"))[0] == 4.5
+
+
+def test_col_by_idx(df_module):
+    assert ns.name(ns.col_by_idx(df_module.example_dataframe, 2)) == "float-col"
+
+
+def test_col_by_idx_duplicate_columns(pd_module):
+    df = pd_module.make_dataframe({"a": [1, 2], "b": [3, 4]})
+    df.columns = ["a", "a"]
+    assert ns.to_list(ns.col_by_idx(df, 0)) == [1, 2]
+
+
 #
 # Querying and modifying metadata
 # ===============================

skrub/_reporting/_summarize.py

@@ -58,7 +58,6 @@
         "n_rows": n_rows,
         "n_columns": n_columns,
         "columns": [],
-        "first_row_dict": _utils.first_row_dict(df) if n_rows else {},
         "dataframe_is_empty": not n_rows or not n_columns,
         "sample_table": _sample_table.make_table(
             df,
@@ -71,17 +70,22 @@
     if order_by is not None:
         df = sbd.sort(df, by=order_by)
         summary["order_by"] = order_by
-    for position, column_name in enumerate(sbd.column_names(df)):
+    if order_by is None:
+        order_by_column = None
+    else:
+        order_by_idx = sbd.column_names(df).index(order_by)
+        order_by_column = sbd.col_by_idx(df, order_by_idx)
+    for position in range(sbd.shape(df)[1]):
         print(
             f"Processing column {position + 1: >3} / {n_columns}", end="\r", flush=True
         )
         summary["columns"].append(
             _summarize_column(
-                sbd.col(df, column_name),
+                sbd.col_by_idx(df, position),
                 position,
                 dataframe_summary=summary,
                 with_plots=with_plots,
-                order_by_column=None if order_by is None else sbd.col(df, order_by),
+                order_by_column=order_by_column,
             )
         )
     print(flush=True)

skrub/_reporting/_table_report.py

@@ -176,7 +176,7 @@
         str :
             The JSON data.
         """
-        to_remove = ["dataframe", "sample_table", "first_row_dict"]
+        to_remove = ["dataframe", "sample_table"]
         data = {
             k: v for k, v in self._summary_without_plots.items() if k not in to_remove
         }

skrub/_reporting/_utils.py

@@ -29,11 +29,6 @@ def _to_dict_polars(df):
     return df.to_dict(as_series=False)
 
 
-def first_row_dict(dataframe):
-    first_row = sbd.slice(dataframe, 0, 1)
-    return {col_name: col[0] for col_name, col in to_dict(first_row).items()}
-
-
 def top_k_value_counts(column, k):
     counts = sbd.value_counts(column)
     n_unique = sbd.shape(counts)[0]

skrub/_reporting/tests/test_summarize.py

@@ -4,7 +4,6 @@
 import numpy as np
 import pandas as pd
 import pytest
-import zoneinfo
 
 from skrub import _column_associations
 from skrub import _dataframe as sbd
@@ -24,21 +23,6 @@ def test_summarize(monkeypatch, df_module, air_quality, order_by, with_plots):
     assert summary["n_columns"] == 11
     assert summary["n_constant_columns"] == 4
     assert summary["n_rows"] == 17
-    assert summary["first_row_dict"] == {
-        "all_null": None,
-        "city": "London",
-        "constant_datetime": datetime.datetime(2024, 7, 5, 12, 17, 29, 427865),
-        "constant_numeric": 2.7,
-        "country": "GB",
-        "date.utc": datetime.datetime(
-            2019, 6, 13, 0, 0, tzinfo=zoneinfo.ZoneInfo(key="UTC")
-        ),
-        "loc_with_nulls": None,
-        "location": "London Westminster",
-        "parameter": "no2",
-        "unit": "µg/m³",
-        "value": 29.0,
-    }
     assert summary["dataframe"] is air_quality
     assert summary["dataframe_module"] == df_module.name
     assert summary["sample_table"]["start_i"] == (
@@ -245,3 +229,15 @@ def test_level_names():
     assert _sample_table._level_names(idx) == ["the name"]
     idx.names = ["a", "b"]
     assert _sample_table._level_names(idx) == ["a", "b"]
+
+
+def test_duplicate_columns(pd_module):
+    df = pd_module.make_dataframe({"a": [1, 2], "b": [3, 4]})
+    df.columns = ["a", "a"]
+    summary = summarize_dataframe(df)
+    cols = summary["columns"]
+    assert len(cols) == 2
+    assert cols[0]["name"] == "a"
+    assert cols[0]["mean"] == 1.5
+    assert cols[1]["name"] == "a"
+    assert cols[1]["mean"] == 3.5

skrub/_reporting/tests/test_table_report.py

@@ -93,3 +93,9 @@ def test_nat(df_module):
     col = ToDatetime().fit_transform(col)
     df = df_module.make_dataframe({"a": col})
     TableReport(df).html()
+
+
+def test_duplicate_columns(pd_module):
+    df = pd_module.make_dataframe({"a": [1, 2], "b": [3, 4]})
+    df.columns = ["a", "a"]
+    TableReport(df).html()