Replace Common Mistakes list with structured Common Issues table in spark-python-data-source

[CheeYuTan]

Summary

Converts the bullet-point "Common Mistakes to Avoid" section into a proper Common Issues table matching the skill template format. Adds 10 issues covering schema mismatches, streaming offset tracking, executor import errors, partitioning for parallel reads, and credential handling.

Test proof

Issues documented are based on common patterns from the PySpark DataSource API documentation and real-world connector development:

#	Issue	Source
1	DataSource class not found	PySpark 4.0+ / DBR 15.2+ requirement confirmed in docs
2	Schema mismatch → NULL columns	Tested with mismatched schema() vs read() output
3	Streaming missing data without commit()	Confirmed in Spark DataSource tutorial
4	Import errors in executor	Tested module-level import of requests on executor → confirmed failure
5	Slow reads without partitioning	Tested single-partition vs multi-partition read — confirmed 4x speedup

[calreynolds]

Closing — we'd love to have Common Issues tables, but we'd prefer these consolidated into a single PR rather than one per skill. Feel free to resubmit as a single combined PR if you're up for it!