[FEATURE] Skipping codegen and compile for Scan only plan

[LantaoJin]

Is your feature request related to a problem?
Currently, when the Calcite engine is enabled, ​​all query plans​​ -- regardless of their structure -- are converted to EnumerableRel. This means the plan is transformed into a Linq4j Expression, generates Java code, and undergoes just-in-time (JIT) compilation before execution.

For ​​simple or fully pushdown-compatible queries​​ (e.g., source=t, source=t | where a=1, source=t | sort a, or source=t | where a=1 | sort b | head 10), the optimized plans end up as EnumerableTableScan. Despite their simplicity, these plans ​​unnecessarily undergo code generation and dynamic compilation​​, introducing overhead.

What solution would you like?
To ​​reduce codegen and compilation time​​, we propose converting such plans to BindableRel instead of EnumerableRel. This bypasses code generation and compilation entirely, ​​improving plan execution time by ~30%​​.

What alternatives have you considered?
A clear and concise description of any alternative solutions or features you've considered.

Do you have any additional context?
Add any other context or screenshots about the feature request here.