Spark job is trigger when compiling SQL with offline engine

Now it takes much time to compile SQL and the data is processed even if getting output dataframe.

The issue mentions a Spark job being triggered during SQL compilation with the offline engine, causing unnecessary data processing and slow compilation. Look at the SQL compilation path in the offline engine code, likely in the Spark-related modules. Identify where the job is being inadvertently triggered and ensure compilation does not execute data processing. Check existing tests for SQL compilation to verify the fix.

Written by the indexing model from the issue text.