release-21.2: opt: fix min cardinality calculation of EXCEPT

[blathers-crl]

Backport 1/1 commits from #89108 on behalf of @rytaft.

/cc @cockroachdb/release

---

Prior to this commit, we assumed that the minimum cardinality of EXCEPT could never be lower than the minimum cardinality of the left side minus the maximum cardinality of the right side. However, this assumption is only true for EXCEPT ALL, not EXCEPT.

For example, VALUES (1), (1) EXCEPT VALUES (1) produces 0 rows, even though the left side produces more rows than the right. Because of this invalid assumption, the optimizer could make some invalid transformations, resulting in incorrect results.

Fixes #89101

Release note (bug fix): Fixed a bug that has existed since v2.1.0 where queries containing a subquery with EXCEPT could produce incorrect results. This could happen if the optimizer could guarantee that the left side of the EXCEPT always returned more rows than the right side. In this case, the optimizer made a faulty assupmtion that the EXCEPT subquery always returned at least one row, which could cause the optimizer to perform an invalid transformation, possibly causing the full query to return incorrect results.

---

Release justification: low risk fix for correctness issue

[blathers-crl]

Thanks for opening a backport.

Please check the backport criteria before merging:

• Patches should only be created for serious issues or test-only changes.
• Patches should not break backwards-compatibility.
• Patches should change as little code as possible.
• Patches should not change on-disk formats or node communication protocols.
• Patches should not add new functionality.
• Patches must not add, edit, or otherwise modify cluster versions; or add version gates.

If some of the basic criteria cannot be satisfied, ensure that the exceptional criteria are satisfied within.

• There is a high priority need for the functionality that cannot wait until the next release and is difficult to address in another way.
• The new functionality is additive-only and only runs for clusters which have specifically “opted in” to it (e.g. by a cluster setting).
• New code is protected by a conditional check that is trivial to verify and ensures that it only runs for opt-in clusters.
• The PM and TL on the team that owns the changed code have signed off that the change obeys the above rules.

Add a brief release justification to the body of your PR to justify this backport.

Some other things to consider:

• What did we do to ensure that a user that doesn’t know & care about this backport, has no idea that it happened?
• Will this work in a cluster of mixed patch versions? Did we test that?
• If a user upgrades a patch version, uses this feature, and then downgrades, what happens?

[cockroach-teamcity]

This change is