DOC: constant check in series (#54064)

* Document constant check in series

Closes #54033

* Update cookbook.rst

* Update cookbook.rst

* empty series is constant

* Update cookbook.rst

* Update cookbook.rst

Author: sbrugman

doc/source/user_guide/cookbook.rst

@@ -1488,3 +1488,31 @@ of the data values:
        {"height": [60, 70], "weight": [100, 140, 180], "sex": ["Male", "Female"]}
    )
    df
+
+Constant series
+---------------
+
+To assess if a series has a constant value, we can check if ``series.nunique() <= 1``.
+However, a more performant approach, that does not count all unique values first, is:
+
+.. ipython:: python
+
+   v = s.to_numpy()
+   is_constant = v.shape[0] == 0 or (s[0] == s).all()
+
+This approach assumes that the series does not contain missing values.
+For the case that we would drop NA values, we can simply remove those values first:
+
+.. ipython:: python
+
+   v = s.dropna().to_numpy()
+   is_constant = v.shape[0] == 0 or (s[0] == s).all()
+
+If missing values are considered distinct from any other value, then one could use:
+
+.. ipython:: python
+
+   v = s.to_numpy()
+   is_constant = v.shape[0] == 0 or (s[0] == s).all() or not pd.notna(v).any()
+
+(Note that this example does not disambiguate between ``np.nan``, ``pd.NA`` and ``None``)