You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@iceberg.apache.org by "fallintoplace (via GitHub)" <gi...@apache.org> on 2026/10/22 08:04:35 UTC
[PR] perf(maintenance): avoid orphan reference index for exact matches [iceberg-go]
fallintoplace opened a new pull request, #2268:
URL: https://github.com/apache/iceberg-go/pull/2268
## What
Skip reference URI index work for exact path matches.
## Why
Most listed files use the same path string as metadata. Building a normalized index for every file adds CPU and allocations.
## Implementation
- Check raw paths first.
- Build the alias index after the first path miss.
- Keep URI aliases and prefix mismatch behavior.
- Add parity tests and paired benchmarks.
## Benchmark
In-memory path matching, Go 1.26.3, Apple M1 Pro, `GOMAXPROCS=1`; 7 runs of 300 ms. Before and after run in the same process with the same inputs. Base: `a4fb4d0`; head: `10cf8a7`. This does not measure object-store listing or full-table cleanup.
| Metric | Before | After | Change |
| --- | ---: | ---: | ---: |
| 100k exact paths, time | 152.894 ms | 2.081 ms | -98.6% |
| 100k exact paths, bytes/op | 70.65 MiB | 0 | -100% |
| 100k exact paths, allocs/op | 800,800 | 0 | -100% |
| 100k files, 1% orphans, time | 152.8 ms | 109.3 ms | -28.5% |
| 100k files, 1% aliases, time | 154.2 ms | 110.6 ms | -28.3% |
| 1% orphans, bytes/op | 70.52 MiB | 49.37 MiB | -30.0% |
| 1% aliases, bytes/op | 70.65 MiB | 49.51 MiB | -29.9% |
Command: `GOMAXPROCS=1 go test ./table/maintenance -run '^$' -bench '^BenchmarkFindOrphanFilesLazyIndex$' -benchmem -benchtime=300ms -count=7 -cpu=1`
--
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: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org
Re: [PR] perf(maintenance): avoid orphan reference index for exact matches [iceberg-go]
Posted by "zeroshade (via GitHub)" <gi...@apache.org>.
zeroshade merged PR #2268:
URL: https://github.com/apache/iceberg-go/pull/2268
--
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: issues-unsubscribe@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org