Fix tpcds planning stack overflows - Join planning refactoring

[Jefffrey]

Which issue does this PR close?

Closes #8837

Rationale for this change

Thanks to the previous analysis from #1047 and with some lldb debugging, was able to identify that create_initial_plan(...) and it's sibling create_initial_plan_multi(...) were the culprits for the stack overflow (also identified by this comment #8837 (comment))

So went looking through it and found a place were there was a large number of local variables before the recursive call to create_initial_plan(...)/create_initial_plan_multi(...) and tested out extracting those into a separate function. This seems to resolve the issue.

What changes are included in this PR?

Extract local variables and functionality in physical join planning in create_initial_plan(...) to a separate function to try reduce stack size before the recursive create_initial_plan(...) call

Are these changes tested?

Are there any user-facing changes?

[alamb]

Thank you @Jefffrey -- this is a great improvement

I tested it locally too. On main:

$ cargo test --test tpcds_planning
...
test tpcds_physical_q5 ... ok

thread 'tpcds_physical_q54' has overflowed its stack
fatal runtime error: stack overflow
error: test failed, to rerun pass `-p datafusion --test tpcds_planning`

On this branch:

$ cargo test --test tpcds_planning

test result: ok. 192 passed; 0 failed; 6 ignored; 0 measured; 0 filtered out; finished in 1.08s

[alamb]

😍

[alamb]

FYI @comphead

[comphead]

wow!