refactor: Remove scan id

[JakubValtar]

This PR removes the scan ID. We rely on caches to ensure that common subplans are executed only once.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.30%. Comparing base (f284700) to head (4c3b2c7).
Report is 6 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #23697   +/-   ##
=======================================
  Coverage   81.29%   81.30%           
=======================================
  Files        1643     1643           
  Lines      223133   223143   +10     
  Branches     2837     2837           
=======================================
+ Hits       181407   181435   +28     
+ Misses      41027    41010   -17     
+ Partials      699      698    -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nameexhaustion]

Could we add an assertion / panic if the ID already exists in the map?

[JakubValtar]

Do you mean to compensate for #23674?

I will add that until it's fixed, even though I would prefer not to do that (it's adding code that compensates for impl details of UniqueId and it's generally hard to find code that depends on impl details).

[JakubValtar]

Actually I don't think I should add that check, it would be quite involved. The scan would probably get inserted first, so I would also have to check for collisions on cache and partitioned sink insertion.

I am working on a fix for the unsoundness, so it should be resolved soon.

Is that acceptable?

[nameexhaustion]

It shouldn't be involved - before cache_nodes.insert(), do assert!(!cache_nodes.contains_key(&id)). That should work, assuming we don't expect duplicate IDs?

[JakubValtar]

I added the checks for duplicate ids.