diff --git a/freqtrade/data/converter/orderflow.py b/freqtrade/data/converter/orderflow.py
index d2df46723..11d17e52f 100644
--- a/freqtrade/data/converter/orderflow.py
+++ b/freqtrade/data/converter/orderflow.py
@@ -115,7 +115,11 @@ def populate_dataframe_with_trades(
 
                 indices = dataframe.index[is_between].tolist()
                 # Add trades to each candle
-                trades_series.loc[indices] = [trades_grouped_df.to_dict(orient="records")]
+                trades_series.loc[indices] = [
+                    trades_grouped_df.drop(columns=["candle_start", "candle_end"]).to_dict(
+                        orient="records"
+                    )
+                ]
                 # Use caching mechanism
                 if (candle_start, candle_next) in cached_grouped_trades:
                     cache_entry = cached_grouped_trades[
@@ -191,7 +195,6 @@ def populate_dataframe_with_trades(
         logger.debug(f"trades.groups_keys in {time.time() - start_time} seconds")
 
         # Merge the complex data Series back into the DataFrame
-        # TODO: maybe candle_start and candle_end should be dropped before assignment?
         dataframe["trades"] = trades_series
         dataframe["orderflow"] = orderflow_series
         dataframe["imbalances"] = imbalances_series
diff --git a/tests/data/test_converter_public_trades.py b/tests/data/test_converter_public_trades.py
index a08ba8d8e..9126edfe3 100644
--- a/tests/data/test_converter_public_trades.py
+++ b/tests/data/test_converter_public_trades.py
@@ -279,6 +279,7 @@ def test_public_trades_trades_mock_populate_dataframe_with_trades__check_trades(
 
     # Assert specific details of the first trade
     t = row["trades"][0]
+    assert list(t.keys()) == ["timestamp", "id", "type", "side", "price", "amount", "cost", "date"]
     assert trades["id"][0] == t["id"]
     assert int(trades["timestamp"][0]) == int(t["timestamp"])
     assert t["side"] == "sell"