1011 Running sql using sparkconnect should not print full stack trace

[b1ackout]

Describe your changes

1. Updated is_non_sqlalchemy_error to handle pyspark.sql.utils.AnalysisException, thus when short_errors is enabled to show only the spark sql error and not the full stack trace

Issue number

Closes #1011

Checklist before requesting a review

• Performed a self-review of my code
• Formatted my code with pkgmt format
• Added tests (when necessary).
• Added docstring documentation and update the changelog (when needed)

---

📚 Documentation preview 📚: https://jupysql--1012.org.readthedocs.build/en/1012/

[b1ackout]

After looking through the code I think adding AnalysisException here will solve the issue since PARSE_SYNTAX_ERROR works as expected.
AnalysisException covers all these error conditions.

[b1ackout]

I just need to test it somehow. Will try to package the jupysql and install it in a spark environment

[edublancas]

you can install like this:

pip install git+https://github.com/b1ackout/jupysql@running-sql-using-sparkconnect-should-not-print-full-stack-trace

[b1ackout]

I tested it but no luck, the error message doesn't contain AnalysisException, only the sql error conditions listed here which are included in AnalysisException. So instead of included all these in the list (which could also be updated regularly) I think checking if the error is of instance of AnalysisException would be a lot cleaner.

Tested it and it works.

[b1ackout]

If AnalysisException is imported then checks if the error is of instance of pyspark's Analysis Exception and handles it accordingly

[b1ackout]

Removed these as they are included in AnalysisException