sql: handle CONSTRAINT...UNIQUE with PARTITION BY ALL / REGIONAL BY ROW

[otan]

sql: teach ADD UNIQUE CONSTRAINT to respect PARTITION ALL BY

Release note (sql change): Fix-up ALTER TABLE ... ADD CONSTRAINT ...
UNIQUE to partition correctly under a PARTITION ALL BY table.

sql: respect REGIONAL BY ROW when adding a new UNIQUE CONSTRAINT

Release note (sql change): This change teaches CRDB to apply zone
configs to new unique constraints in REGIONAL BY ROW tables.

[cockroach-teamcity]

This change is 

[ajstorm]

Just one nit below.

Reviewed 2 of 2 files at r1, 2 of 2 files at r2.
Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @otan)

---

pkg/ccl/logictestccl/testdata/logic_test/regional_by_row, line 207 at r2 (raw file):

statement ok
ALTER TABLE regional_by_row_table ADD CONSTRAINT unique_b_a UNIQUE(b, a)

Nit: We're testing adding a unique constraint on ALTER. Can we also test adding one on CREATE?

[otan]

---

pkg/ccl/logictestccl/testdata/logic_test/regional_by_row, line 207 at r2 (raw file):

Previously, ajstorm wrote…

Nit: We're testing adding a unique constraint on ALTER. Can we also test adding one on CREATE?

There already is one on (b)

[otan]

thanks

bors r=ajstorm

[craig]

Build succeeded:

• GitHub CI (Cockroach)

[ajstorm]

Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @otan)

---

pkg/ccl/logictestccl/testdata/logic_test/regional_by_row, line 207 at r2 (raw file):

Previously, otan (Oliver Tan) wrote…

There already is one on (b)

Isn't there just a unique index on b? Are unique constraints and unique indexes handle identically under the covers?

[otan]

---

pkg/ccl/logictestccl/testdata/logic_test/regional_by_row, line 207 at r2 (raw file):

Previously, ajstorm wrote…

Isn't there just a unique index on b? Are unique constraints and unique indexes handle identically under the covers?

in CRDB they are treated the same (as of writing)