release-21.1: opt: fix bug in histogram estimation code for multi-column spans

[rytaft]

Backport 1/1 commits from #76486.

/cc @cockroachdb/release

Release justification: Low risk, high benefit change to existing functionality

---

This commit fixes a bug in the histogram estimation code, which could
cause the optimizer to think that an index scan produced 0 rows, when
in fact it produced a large number. This was due to an inaccurate assumption
in the histogram filtering code that if a span had an exclusive boundary,
the upper bound of the span was excluded from the histogram. However, this
failed to account for the fact that we support constraining a histogram with
multi-column spans, and we can select different column offsets to use to
constrain the histogram. The assumption above is only valid if the column
offset corresponds to the last column in the span key. This logic has now
been fixed.

Fixes #76485

Release note (performance improvement): Fixed a bug in the histogram estimation
code that could cause the optimizer to think a scan of a multi-column index
would produce 0 rows, when in fact it would produce many rows. This could cause
the optimizer to choose a suboptimal plan. This bug has now been fixed, making
it less likely for the optimizer to choose a suboptimal plan when multiple
multi-column indexes are available.

[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 

[mgartner]

LGTM