[SPARK-33017] PySpark Context should have getCheckpointDir() method

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 3.1.0
• Fix Version/s: 3.1.0
• Component/s: PySpark
• Assignee: Paul Reidy
• Reporter: Nicholas Chammas
• Created: 28/Sep/20 14:56
• Updated: 12/Dec/22 18:10
• Resolved: 05/Oct/20 02:49

Description

To match the Scala API, PySpark should offer a direct way to get the checkpoint dir.

scala> spark.sparkContext.setCheckpointDir("/tmp/spark/checkpoint")
scala> spark.sparkContext.getCheckpointDir
res3: Option[String] = Some(file:/tmp/spark/checkpoint/34ebe699-bc83-4c5d-bfa2-50451296cf87)

Currently, the only was to do that from PySpark is via the underlying Java context:

>>> spark.sparkContext.setCheckpointDir('/tmp/spark/checkpoint/')
>>> sc._jsc.sc().getCheckpointDir().get()
'file:/tmp/spark/checkpoint/ebf0fab5-edbc-42c2-938f-65d5e599cf54'

Issue Links

• links to: [Github] Pull Request #29918 (reidy-p)
• links to: [Github] Pull Request #29960 (HyukjinKwon)