diff --git a/python/cudf/cudf/core/column/column.py b/python/cudf/cudf/core/column/column.py
index 1a83194489d..7999fa9039b 100644
--- a/python/cudf/cudf/core/column/column.py
+++ b/python/cudf/cudf/core/column/column.py
@@ -1213,11 +1213,7 @@ def corr(self, other: ColumnBase):
         )
 
     def nans_to_nulls(self: T) -> T:
-        # Only floats can contain nan.
-        if self.dtype.kind != "f":
-            return self
-        newmask = libcudf.transform.nans_to_nulls(self)
-        return self.set_mask(newmask)
+        return self
 
     def _process_for_reduction(
         self, skipna: bool = None, min_count: int = 0
diff --git a/python/cudf/cudf/core/column/numerical.py b/python/cudf/cudf/core/column/numerical.py
index 2b0d7cfea38..a7481ce62a3 100644
--- a/python/cudf/cudf/core/column/numerical.py
+++ b/python/cudf/cudf/core/column/numerical.py
@@ -217,6 +217,13 @@ def binary_operator(
         lhs, rhs = (self, rhs) if not reflect else (rhs, self)
         return libcudf.binaryop.binaryop(lhs, rhs, binop, out_dtype)
 
+    def nans_to_nulls(self: NumericalColumn) -> NumericalColumn:
+        # Only floats can contain nan.
+        if self.dtype.kind != "f" or self.nan_count == 0:
+            return self
+        newmask = libcudf.transform.nans_to_nulls(self)
+        return self.set_mask(newmask)
+
     def normalize_binop_value(
         self, other: ScalarLike
     ) -> Union[ColumnBase, ScalarLike]: