diff --git a/src/synthesized_datasets/_datasets.py b/src/synthesized_datasets/_datasets.py
index 3057bb4..0a4f06c 100644
--- a/src/synthesized_datasets/_datasets.py
+++ b/src/synthesized_datasets/_datasets.py
@@ -74,22 +74,19 @@ def load(self) -> _pd.DataFrame:
         else:
             # CSV load is the default
             dtypes = {
-                col: (
-                    _PD_DTYPE_MAP[dtype]
-                    if dtype
-                    not in [
-                        _DType.DATETIME,
-                        _DType.TIMEDELTA,
-                        _DType.DATE,
-                        _DType.TIME,
-                    ]
-                    else "string"
-                )
+                col: _PD_DTYPE_MAP[dtype]
                 for col, dtype in self._schema.items()
+                if dtype
+                not in [
+                    _DType.DATETIME,
+                    _DType.TIMEDELTA,
+                    _DType.DATE,
+                    _DType.TIME,
+                ]
             }
             df = _pd.read_csv(self.url, dtype=dtypes)
             for col, dtype in self._schema.items():
-                if dtype is [_DType.DATETIME, _DType.DATE]:
+                if dtype in [_DType.DATETIME, _DType.DATE]:
                     df[col] = _pd.to_datetime(df[col], dayfirst=True)
                 if dtype in [_DType.TIMEDELTA, _DType.TIME]:
                     df[col] = _pd.to_timedelta(df[col])