From aa619ec423c63678852dd3765a1ee7755e59cba3 Mon Sep 17 00:00:00 2001
From: cngmid <edu@cngmi.com>
Date: Wed, 20 Aug 2025 14:43:35 +0200
Subject: [PATCH 1/2] Fixing bug #1930 with proper syntax

---
 .../data/timeseries/_timeseries.py            | 11 ++++--
 tests/test_data/test_timeseries.py            | 38 +++++++++++++++++++
 2 files changed, 46 insertions(+), 3 deletions(-)

diff --git a/pytorch_forecasting/data/timeseries/_timeseries.py b/pytorch_forecasting/data/timeseries/_timeseries.py
index 98d16c920..1a9a4b586 100644
--- a/pytorch_forecasting/data/timeseries/_timeseries.py
+++ b/pytorch_forecasting/data/timeseries/_timeseries.py
@@ -1430,7 +1430,7 @@ def _data_to_tensors(self, data: pd.DataFrame) -> dict[str, torch.Tensor]:
             time index
         """
 
-        def _to_tensor(cols, long=True) -> torch.Tensor:
+        def _to_tensor(cols, long=True, real=False) -> torch.Tensor:
             """Convert data[cols] to torch tensor.
 
             Converts sub-frames to numpy and then to torch tensor.
@@ -1439,16 +1439,21 @@ def _to_tensor(cols, long=True) -> torch.Tensor:
             * float columns are converted to torch.float
             * integer columns are converted to torch.int64 or torch.long,
               depending on the long argument
+            * mixed columns are converted to their commonest type.
+            * if real argument is True, force the conversion to torch.float
             """
             if not isinstance(cols, list) and cols not in data.columns:
                 return None
             if isinstance(cols, list) and len(cols) == 0:
                 dtypekind = "f"
             elif isinstance(cols, list):  # and len(cols) > 0
-                dtypekind = data.dtypes[cols[0]].kind
+                # dtypekind = data.dtypes[cols[0]].kind
+                dtypekind = data[cols].to_numpy().dtype.kind
             else:
                 dtypekind = data.dtypes[cols].kind
-            if not long:
+            if real:
+                return torch.tensor(data[cols].to_numpy(np.float64), dtype=torch.float)
+            elif not long:
                 return torch.tensor(data[cols].to_numpy(np.int64), dtype=torch.int64)
             elif dtypekind in "bi":
                 return torch.tensor(data[cols].to_numpy(np.int64), dtype=torch.long)
diff --git a/tests/test_data/test_timeseries.py b/tests/test_data/test_timeseries.py
index 0b1b0ce74..4bb909792 100644
--- a/tests/test_data/test_timeseries.py
+++ b/tests/test_data/test_timeseries.py
@@ -678,3 +678,41 @@ def distance_to_weights(dist):
         if idx > 100:
             break
     print(a)
+
+
+def test_correct_dtype_inference():
+    # Create a small dataset
+    data = pd.DataFrame(
+        {
+            "time_idx": np.arange(30),
+            "value": np.sin(np.arange(30) / 5) + np.random.normal(scale=1, size=30),
+            "group": ["A"] * 30,
+        }
+    )
+
+    # Define the dataset
+    dataset = TimeSeriesDataSet(
+        data.copy(),
+        time_idx="time_idx",
+        target="value",
+        group_ids=["group"],
+        static_categoricals=["group"],
+        max_encoder_length=4,
+        max_prediction_length=2,
+        time_varying_unknown_reals=["value"],
+        target_normalizer=None,
+        # WATCH THIS
+        time_varying_known_reals=["time_idx"],
+        scalers=dict(time_idx=None),
+    )
+
+    # and the dataloader
+    dataloader = dataset.to_dataloader(batch_size=8)
+
+    x, y = next(iter(dataset))
+    # real features must be real
+    assert x["x_cont"].dtype is torch.float32
+
+    x, y = next(iter(dataloader))
+    # real features must be real
+    assert x["encoder_cont"].dtype is torch.float32

From 9d2f5b1deeabea170b3dfa7ada14388b75c7a6a8 Mon Sep 17 00:00:00 2001
From: cngmid <edu@cngmi.com>
Date: Tue, 26 Aug 2025 09:00:11 +0200
Subject: [PATCH 2/2] Modify fix for bug #1930: updated _to_tensor logic

---
 pytorch_forecasting/data/timeseries/_timeseries.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pytorch_forecasting/data/timeseries/_timeseries.py b/pytorch_forecasting/data/timeseries/_timeseries.py
index 1a9a4b586..d4c812d13 100644
--- a/pytorch_forecasting/data/timeseries/_timeseries.py
+++ b/pytorch_forecasting/data/timeseries/_timeseries.py
@@ -1448,7 +1448,7 @@ def _to_tensor(cols, long=True, real=False) -> torch.Tensor:
                 dtypekind = "f"
             elif isinstance(cols, list):  # and len(cols) > 0
                 # dtypekind = data.dtypes[cols[0]].kind
-                dtypekind = data[cols].to_numpy().dtype.kind
+                dtypekind = np.result_type(*data[cols].dtypes.to_list()).kind
             else:
                 dtypekind = data.dtypes[cols].kind
             if real: