Baymine opened a new pull request, #66363:
URL: https://github.com/apache/doris/pull/66363
### What problem does this PR solve?
Issue Number: close #66362
Related PR: None
Problem Summary:
JSON path extraction eagerly formats an error message at every
`HANDLE_SIMDJSON_ERROR` call before checking whether simdjson reported an
error. Successful lookups therefore pay repeated formatting and allocation
costs even though the message is unused.
This change constructs the message only inside the unlikely error branch. It
keeps owned string storage on that branch so formatted temporaries remain safe.
Focused coverage verifies missing fields, out-of-bounds array indexes, and
incorrect intermediate types, and a benchmark covers successful and
missing-field lookups.
In a local same-session RELEASE A/B benchmark over 4,096 rows, the
valid-field median improved from 689,024 ns to 364,504 ns (about 1.89x
throughput); the missing-field path remained approximately unchanged.
### Release note
None
### Check List (For Author)
- Test
- [ ] Regression test
- [x] Unit Test
- [x] Manual test: `./build.sh --be`, `./build.sh --benchmark`, and the
A/B benchmark described above
- [ ] No need to test or manual test.
- Behavior changed:
- [x] No.
- [ ] Yes.
- Does this need documentation?
- [x] No.
- [ ] Yes.
### Check List (For Reviewer who merge this PR)
- [ ] Confirm the release note
- [ ] Confirm test cases
- [ ] Confirm document
- [ ] Add branch pick label
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
