added documentation in docstring for the specific case where orient=table and column is a numeric value and not string

Author: chandra-teajunkie

pandas/core/generic.py

@@ -2377,6 +2377,9 @@ def to_json(
 
                 Describing the data, where data component is like ``orient='records'``.
 
+                Note that If `orient='table'`, column names must be strings.
+                Using numeric column names will raise a `ValueError`.
+
         date_format : {{None, 'epoch', 'iso'}}
             Type of date conversion. 'epoch' = epoch milliseconds,
             'iso' = ISO8601. The default depends on the `orient`. For
@@ -2446,6 +2449,10 @@ def to_json(
         This stores the version of `pandas` used in the latest revision of the
         schema.
 
+        When using ``orient='table'``, column names must be strings due to JSON
+        requiring string keys. If column names are numeric (integers or floats),
+        a `ValueError` will be raised.
+
         Examples
         --------
         >>> from json import loads, dumps