Match Phase Fails Without Explicit Checkpoint in Zingg 0.5.0 on Databricks

[padam-prakash]

Describe the bug

When running Zingg version 0.5.0 on Databricks, the Match phase completes but does not produce any output unless a checkpoint directory is explicitly set.

---

Steps to reproduce the issue

1. Run a Databricks notebook using the Zingg 0.5.0 JAR.
2. Do not set a checkpoint directory before executing the Match phase.
3. Execute the Match phase.
   • Result: The Match phase completes but no output is generated.
4. Set a checkpoint directory.
5. Re-run the Match phase.
   • Result: The Match phase succeeds and the output is generated correctly.

---

Expected behavior

The Match phase should run successfully and output should be generated correctly even if the checkpoint directory is not explicitly set by the user.

---

Environment details

• Platform: Databricks
• Databricks Runtime Version: 16.4 LTS (includes Apache Spark 3.5.2, Scala 2.12)
• Zingg Version: 0.5.0
• Cluster Type: Single node

---

[padam-prakash]

#1207