Description
pidx = pd.Index([10, 20, 15, 30, 45, None], name="x") psidx = ps.Index(pidx) self.assert_eq(psidx.astype(str), pidx.astype(str))
[left pandas on spark]: Index(['10.0', '20.0', '15.0', '30.0', '45.0', 'nan'], dtype='object', name='x')
[right pandas]: Index(['10', '20', '15', '30', '45', 'None'], dtype='object', name='x')
The index is loaded as float64, so the follow step like astype would be diff with pandas
Attachments
Issue Links
- relates to
-
SPARK-34849 SPIP: Support pandas API layer on PySpark
- Resolved
- links to