pyspark.pandas.Index.is_numeric

Index.is_numeric() → bool[source]

Return if the current index type is a numeric type.

Examples

>>> ps.DataFrame({'a': [1]}, index=[1]).index.is_numeric()
True