DOC: clarify idxmax behaviour issue #18206 (#18209)

(cherry picked from commit 63e8527)

Author: ghasemnaddaf

pandas/core/frame.py

@@ -5813,7 +5813,12 @@ def idxmin(self, axis=0, skipna=True):
             0 or 'index' for row-wise, 1 or 'columns' for column-wise
         skipna : boolean, default True
             Exclude NA/null values. If an entire row/column is NA, the result
-            will be NA
+            will be NA.
+
+        Raises
+        ------
+        ValueError
+            * If the row/column is empty
 
         Returns
         -------
@@ -5844,7 +5849,12 @@ def idxmax(self, axis=0, skipna=True):
             0 or 'index' for row-wise, 1 or 'columns' for column-wise
         skipna : boolean, default True
             Exclude NA/null values. If an entire row/column is NA, the result
-            will be first index.
+            will be NA.
+
+        Raises
+        ------
+        ValueError
+            * If the row/column is empty
 
         Returns
         -------

pandas/core/series.py

@@ -1306,7 +1306,13 @@ def idxmin(self, axis=None, skipna=True, *args, **kwargs):
         Parameters
         ----------
         skipna : boolean, default True
-            Exclude NA/null values
+            Exclude NA/null values. If the entire Series is NA, the result
+            will be NA.
+
+        Raises
+        ------
+        ValueError
+            * If the Series is empty
 
         Returns
         -------
@@ -1336,7 +1342,13 @@ def idxmax(self, axis=None, skipna=True, *args, **kwargs):
         Parameters
         ----------
         skipna : boolean, default True
-            Exclude NA/null values
+            Exclude NA/null values. If the entire Series is NA, the result
+            will be NA.
+
+        Raises
+        ------
+        ValueError
+            * If the Series is empty
 
         Returns
         -------