From 33803db2cb9befbfd5fef11f81a8bf2ebd014185 Mon Sep 17 00:00:00 2001
From: arcwiser <arcwiser@users.noreply.github.com>
Date: Tue, 13 Oct 2026 01:07:15 +1100
Subject: [PATCH] Change list_setups default to dataframe

Signed-off-by: arcwiser <arcwiser@users.noreply.github.com>
---
 openml/setups/functions.py                | 10 +++++--
 tests/test_setups/test_setup_functions.py | 35 ++++++++++++++++-------
 2 files changed, 32 insertions(+), 13 deletions(-)

diff --git a/openml/setups/functions.py b/openml/setups/functions.py
index 0b1a4f375..13bd2c6d9 100644
--- a/openml/setups/functions.py
+++ b/openml/setups/functions.py
@@ -77,7 +77,7 @@ def list_setups(  # noqa: PLR0913
     flow: int | None = None,
     tag: str | None = None,
     setup: Iterable[int] | None = None,
-    output_format: Literal["object", "dataframe"] = "object",
+    output_format: Literal["object", "dataframe"] = "dataframe",
 ) -> dict[int, OpenMLSetup] | pd.DataFrame:
     """
     List all setups matching all of the given filters.
@@ -89,7 +89,7 @@ def list_setups(  # noqa: PLR0913
     flow : int, optional
     tag : str, optional
     setup : Iterable[int], optional
-    output_format: str, optional (default='object')
+    output_format: str, optional (default='dataframe')
         The parameter decides the format of the output.
         - If 'dataframe' the output is a pandas DataFrame
         - If 'object' the output is a dictionary of OpenMLSetup objects
@@ -115,7 +115,11 @@ def list_setups(  # noqa: PLR0913
         return {setup.setup_id: setup for setup in flattened}
 
     records = [setup._to_dict() for setup in flattened]
-    return pd.DataFrame.from_records(records, index="setup_id")
+    return pd.DataFrame.from_records(
+        records,
+        columns=["setup_id", "flow_id", "parameters"],
+        index="setup_id",
+    )
 
 
 def initialize_model(setup_id: int, *, strict_version: bool = True) -> Any:
diff --git a/tests/test_setups/test_setup_functions.py b/tests/test_setups/test_setup_functions.py
index 0735925f2..b0f18c12a 100644
--- a/tests/test_setups/test_setup_functions.py
+++ b/tests/test_setups/test_setup_functions.py
@@ -17,6 +17,21 @@
 from openml.testing import TestBase
 
 
+@pytest.mark.parametrize(
+    ("batches", "expected_index"),
+    [([[openml.setups.setup.OpenMLSetup(7, 11, None)]], [7]), ([], [])],
+)
+def test_list_setups_defaults_to_dataframe(mocker, batches, expected_index):
+    mocker.patch("openml.utils._list_all", return_value=batches)
+
+    setups = openml.setups.list_setups()
+
+    assert isinstance(setups, pd.DataFrame)
+    assert setups.index.name == "setup_id"
+    assert list(setups.index) == expected_index
+    assert list(setups.columns) == ["flow_id", "parameters"]
+
+
 def get_sentinel():
     # Create a unique prefix for the flow. Necessary because the flow is
     # identified by its name and external version online. Having a unique
@@ -141,7 +156,7 @@ def test_setup_list_filter_flow(self):
 
         flow_id = 5873
 
-        setups = openml.setups.list_setups(flow=flow_id)
+        setups = openml.setups.list_setups(flow=flow_id, output_format="object")
 
         assert len(setups) >= 2
         for setup_id in setups:
@@ -150,22 +165,22 @@ def test_setup_list_filter_flow(self):
     @pytest.mark.test_server()
     def test_list_setups_empty(self):
         setups = openml.setups.list_setups(setup=[0])
-        if len(setups) > 0:
-            raise ValueError("UnitTest Outdated, got somehow results")
-
-        assert isinstance(setups, dict)
+        assert isinstance(setups, pd.DataFrame)
+        assert setups.empty
+        assert setups.index.name == "setup_id"
+        assert list(setups.columns) == ["flow_id", "parameters"]
 
     @pytest.mark.production_server()
     def test_list_setups_output_format(self):
         self.use_production_server()
         flow_id = 6794
         setups = openml.setups.list_setups(flow=flow_id, size=10)
-        assert isinstance(setups, dict)
-        assert isinstance(setups[next(iter(setups.keys()))], openml.setups.setup.OpenMLSetup)
+        assert isinstance(setups, pd.DataFrame)
         assert len(setups) == 10
 
-        setups = openml.setups.list_setups(flow=flow_id, size=10, output_format="dataframe")
-        assert isinstance(setups, pd.DataFrame)
+        setups = openml.setups.list_setups(flow=flow_id, size=10, output_format="object")
+        assert isinstance(setups, dict)
+        assert isinstance(setups[next(iter(setups))], openml.setups.setup.OpenMLSetup)
         assert len(setups) == 10
 
     @pytest.mark.test_server()
@@ -176,7 +191,7 @@ def test_setuplist_offset(self):
         setups2 = openml.setups.list_setups(offset=size, size=size)
         assert len(setups2) == size
 
-        all = set(setups.keys()).union(setups2.keys())
+        all = set(setups.index).union(setups2.index)
 
         assert len(all) == size * 2