sql/jsonpath: support index acceleration with AnyKey (*) at the chain end

[ZhouXing19]

fixes: #156340

We now support index accelerating jsonb_path_exists filters with json path expression that ends with an AnyKey (*).

Note that the AnyKey is allowed only at the end of the expression. I.e. the following are not allowed:

$.a.*.b 
$.a.b.*.* 

Release note (sql change): We now support index accelerating jsonb_path_exists filters with json path expression that ends with an AnyKey (*).

[cockroach-teamcity]

This change is 

[yuzefovich]

Nice, looks good to me, just some nits.

@yuzefovich reviewed 5 of 5 files at r1, all commit messages.
Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @ZhouXing19)

---

pkg/util/jsonpath/path.go line 319 at r1 (raw file):

// isSupportedPathPattern returns true if the given paths matches one of
// the following patterns, which can be supported by the inverted index:
// - keychain mode: $.[key|wildcard].[key|wildcard]...

nit: let's update the comment to include optional AnyKey at the end.

---

pkg/util/jsonpath/path.go line 592 at r1 (raw file):

						// In these 2 cases, after "b", there is still "c"
						// as the next key, so "b" is not an end key.
						if _, isAnyKey := pathComponents[j].(AnyKey); isAnyKey {

Given the check in isSupportedPathPattern, isAnyKey can only be true when j == len(pathComponents)-1, right? Perhaps we should only check that specific value of j?

---

pkg/sql/logictest/testdata/logic_test/jsonb_path_exists_index_acceleration line 560 at r1 (raw file):

SELECT a FROM json_tab@foo_inv WHERE jsonb_path_exists(b, '$.a ? (@.b == $x)', '{"x": "c"}') ORDER BY a;

subtest anykey

nit: we usually add an empty line after subtest directives.

[ZhouXing19]

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

---

pkg/util/jsonpath/path.go line 319 at r1 (raw file):

Previously, yuzefovich (Yahor Yuzefovich) wrote…

nit: let's update the comment to include optional AnyKey at the end.

Done.

---

pkg/util/jsonpath/path.go line 592 at r1 (raw file):

Previously, yuzefovich (Yahor Yuzefovich) wrote…

Given the check in isSupportedPathPattern, isAnyKey can only be true when j == len(pathComponents)-1, right? Perhaps we should only check that specific value of j?

Yes, good idea! Done.

[yuzefovich]

I added the backport label - it seems like a nice improvement to the new feature with low risk, let's see whether Michael agrees.

@yuzefovich reviewed 2 of 2 files at r2, all commit messages.
Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @ZhouXing19)

[ZhouXing19]

TFTR!
bors r+

[craig]

This PR was included in a batch that successfully built, but then failed to merge into master (it was a non-fast-forward update). It will be automatically retried.

[craig]

Build succeeded:

• linux_amd64_build
• acceptance
• linux_amd64_fips_build
• docker_image_amd64
• local_roachtest_fips
• unit_tests
• examples_orms
• lint
• check_generated_code
• local_roachtest

[blathers-crl]

Based on the specified backports for this PR, I applied new labels to the following linked issue(s). Please adjust the labels as needed to match the branches actually affected by the issue(s), including adding any known older branches.

---

Issue #156340: branch-release-25.4.

---

_{🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf.}