feat[file]: gate the writer on the session's enabled editions#8946
Open
joseph-isaacs wants to merge 5 commits into
Open
feat[file]: gate the writer on the session's enabled editions#8946joseph-isaacs wants to merge 5 commits into
joseph-isaacs wants to merge 5 commits into
Conversation
Adds a required `Scheme::produced_encodings` method returning the array encodings a scheme may itself introduce into its compressed output. Cascaded children are compressed by other schemes, which declare their own encodings, so only encodings constructed directly by `compress` belong here. `BtrBlocksCompressorBuilder::retain_allowed_encodings` uses the declarations to drop any scheme that could produce an encoding outside a permitted set, so an encoding policy can be enforced before compression rather than rejected after it. Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Introduces `EnabledEditions`, a session variable recording the editions a writer is permitted to emit — at most one per family, with a newer or older edition from the same family replacing the previous selection. This is deliberately separate from `EditionSession`: registration describes what a session knows how to reason about, while enabling is the explicit writer policy. `EditionSessionExt` gains: - `enabled_editions` — access the (lazily installed) selection. - `register_edition` / `enable_edition` — enabling an unregistered edition errors, so a typo or unavailable third-party declaration cannot silently produce an empty writable set. - `enabled_encoding_ids` — resolve the union of encodings across all enabled editions; empty when nothing is enabled. The default `vortex` session now enables the newest frozen `core` edition (and the newest unstable edition under `unstable_encodings`) via `editions::enable_default_editions`. No writer behavior changes yet: this only records the selection on the session. A follow-up PR consumes `enabled_encoding_ids` to gate the file writer. Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
The file writer previously drew its permitted encodings from a static `ALLOWED_ENCODINGS` set in `vortex-file`. That set was compiled in, so the only way to write a narrower or wider encoding set was to override the write strategy by hand, and the writer's policy was disconnected from the editions the session had registered. Encoding policy now comes from the session's enabled editions: - `VortexWriteOptions::new` resolves `enabled_encoding_ids` and applies it both to the write strategy and to the `ArrayContext`, so a replacement strategy cannot widen the policy. - `WriteStrategyBuilder::with_allow_encodings` now also filters the BtrBlocks compressor via `retain_allowed_encodings`, in either configuration order, so forbidden encodings are never produced rather than being rejected after the fact. - `ALLOWED_ENCODINGS` is removed. `register_default_encodings` still registers everything for reading; the default session's write policy comes from the editions enabled by `enable_default_editions`. Sessions without an `EnabledEditions` variable permit no edition encodings, so lower-level test sessions in `vortex-file`, `vortex-layout` and `vortex-parquet-variant` opt in explicitly (a shared `enable_all_registered_array_encodings` test helper). Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
Move `retain_allowed_encodings` out of the `with_allow_encodings` and `with_btrblocks_builder` setters and into `build()`, applying it once to the compressor before it is used for the data, stats, and defaulted probe compressors. This drops the order-dependence between configuring the builder and the allowed-encodings policy: whichever order they are set, the filter is applied to every compressor derived from the builder. Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>
joseph-isaacs
changed the base branch from
ji/enabled-editions-session
to
develop
July 24, 2026 14:54
joseph-isaacs
added
the
action/benchmark
github-actions
Bot
removed
the
action/benchmark
…tions Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> # Conflicts: # vortex-jni/src/writer.rs # vortex-layout/src/layouts/struct_/reader.rs # vortex-layout/src/layouts/table.rs # vortex-layout/src/test.rs
Contributor
Polar Signals Profiling ResultsLatest Run
Powered by Polar Signals Cloud |
Contributor
Benchmarks: PolarSignals Profiling 📖Vortex (geomean): 1.009x ➖ How to read Verdict and Engines
datafusion / vortex-file-compressed (1.009x ➖, 0↑ 0↓)
File Size Changes (1 files changed, +0.0% overall, 1↑ 0↓)
Totals:
|
Contributor
Benchmarks: FineWeb NVMe 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.031x ➖, 0↑ 1↓)
datafusion / vortex-compact (1.035x ➖, 0↑ 1↓)
datafusion / parquet (1.004x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.059x ➖, 0↑ 2↓)
duckdb / vortex-compact (1.011x ➖, 0↑ 0↓)
duckdb / parquet (0.997x ➖, 1↑ 0↓)
File Size Changes (2 files changed, +0.0% overall, 2↑ 0↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=1 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.001x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.007x ➖, 0↑ 0↓)
datafusion / parquet (1.026x ➖, 0↑ 2↓)
datafusion / arrow (1.013x ➖, 0↑ 1↓)
duckdb / vortex-file-compressed (1.030x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.027x ➖, 0↑ 0↓)
duckdb / parquet (1.004x ➖, 0↑ 1↓)
duckdb / duckdb (1.029x ➖, 0↑ 0↓)
File Size Changes (18 files changed, +0.0% overall, 18↑ 0↓)
Totals:
|
Contributor
Benchmarks: Vortex queries 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.844x ✅, 2↑ 0↓)
datafusion / parquet (0.886x ✅, 2↑ 0↓)
duckdb / vortex-file-compressed (0.916x ➖, 1↑ 0↓)
duckdb / parquet (0.938x ➖, 0↑ 0↓)
File Size Changes (1 files changed, +0.0% overall, 1↑ 0↓)
Totals:
|
Contributor
Benchmarks: TPC-DS SF=1 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.039x ➖, 0↑ 7↓)
datafusion / vortex-compact (1.037x ➖, 0↑ 8↓)
datafusion / parquet (1.036x ➖, 0↑ 5↓)
duckdb / vortex-file-compressed (1.030x ➖, 0↑ 7↓)
duckdb / vortex-compact (1.038x ➖, 1↑ 6↓)
duckdb / parquet (1.007x ➖, 1↑ 0↓)
duckdb / duckdb (1.038x ➖, 0↑ 3↓)
File Size Changes (48 files changed, +0.0% overall, 48↑ 0↓)
Totals:
|
Contributor
Benchmarks: FineWeb S3 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.001x ➖, 0↑ 0↓)
datafusion / vortex-compact (0.980x ➖, 0↑ 0↓)
datafusion / parquet (1.042x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (1.069x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.065x ➖, 0↑ 0↓)
duckdb / parquet (1.017x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: Statistical and Population Genetics 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
duckdb / vortex-file-compressed (0.996x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.998x ➖, 0↑ 0↓)
duckdb / parquet (0.993x ➖, 0↑ 0↓)
File Size Changes (2 files changed, +0.0% overall, 2↑ 0↓)
Totals:
|
Contributor
Benchmarks: Clickbench Sorted on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.006x ➖, 1↑ 2↓)
datafusion / parquet (0.950x ➖, 1↑ 0↓)
duckdb / vortex-file-compressed (0.991x ➖, 1↑ 1↓)
duckdb / parquet (0.984x ➖, 0↑ 0↓)
duckdb / duckdb (0.973x ➖, 0↑ 0↓)
File Size Changes (201 files changed, +0.0% overall, 100↑ 101↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=10 on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.757x ✅, 22↑ 0↓)
datafusion / vortex-compact (0.770x ✅, 22↑ 0↓)
datafusion / parquet (0.801x ✅, 21↑ 0↓)
datafusion / arrow (0.763x ✅, 22↑ 0↓)
duckdb / vortex-file-compressed (0.803x ✅, 22↑ 0↓)
duckdb / vortex-compact (0.819x ✅, 22↑ 0↓)
duckdb / parquet (0.893x ✅, 15↑ 0↓)
duckdb / duckdb (0.864x ✅, 18↑ 0↓)
File Size Changes (48 files changed, +0.0% overall, 48↑ 0↓)
Totals:
|
Contributor
Benchmarks: Clickbench on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.013x ➖, 0↑ 1↓)
datafusion / parquet (1.004x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.991x ➖, 6↑ 1↓)
duckdb / parquet (0.996x ➖, 1↑ 0↓)
duckdb / duckdb (1.002x ➖, 0↑ 0↓)
File Size Changes (201 files changed, -0.0% overall, 200↑ 1↓)
Totals:
|
Contributor
Benchmarks: Random Access 📖Vortex (geomean): 0.964x ➖ How to read Verdict and Engines
unknown / unknown (0.994x ➖, 1↑ 1↓)
|
Contributor
Benchmarks: TPC-H SF=1 on S3 📖Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.005x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.029x ➖, 0↑ 2↓)
datafusion / parquet (1.046x ➖, 0↑ 1↓)
duckdb / vortex-file-compressed (1.027x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.014x ➖, 0↑ 0↓)
duckdb / parquet (1.017x ➖, 0↑ 0↓)
|
joseph-isaacs
force-pushed
the
ji/gate-writer-editions
branch
from
July 24, 2026 15:17
4a21a9f to
16b2711
Compare
Contributor
Benchmarks: Appian on NVME 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.994x ➖, 0↑ 0↓)
datafusion / parquet (0.995x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.989x ➖, 0↑ 0↓)
duckdb / parquet (0.994x ➖, 0↑ 0↓)
duckdb / duckdb (0.995x ➖, 0↑ 0↓)
File Size Changes (19 files changed, -0.0% overall, 18↑ 1↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=10 on S3 📖Verdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.011x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.022x ➖, 0↑ 0↓)
datafusion / parquet (0.965x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.959x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.934x ➖, 0↑ 0↓)
duckdb / parquet (0.994x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: Compression 📖Vortex (geomean): 1.000x ➖ How to read Verdict and Engines
unknown / unknown (0.999x ➖, 0↑ 0↓)
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The file writer previously drew its permitted encodings from a static
ALLOWED_ENCODINGSset compiled intovortex-file. The only way to write a narrower or wider set was to override the write strategy by hand, and the writer's policy was disconnected from the editions the session had registered. This PR makes the writer's encoding policy come from the session's enabled editions.Final PR in the stack, on top of #8945 (
add enabled-editions selection to the session) and #8944 (require schemes to declare their produced encodings).Changes
VortexWriteOptions::newresolvesenabled_encoding_idsand applies it both to the write strategy and to theArrayContext, so a replacement strategy cannot widen the policy.WriteStrategyBuilder::with_allow_encodingsnow also filters the BtrBlocks compressor viaretain_allowed_encodings(from feat[compressor]: require schemes to declare their produced encodings #8944), in either configuration order, so forbidden encodings are never produced rather than being rejected after the fact.ALLOWED_ENCODINGSis removed.register_default_encodingsstill registers everything for reading; the default session's write policy comes from the editions enabled byenable_default_editions.EnabledEditionsvariable permit no edition encodings, so lower-level test sessions invortex-file,vortex-layoutandvortex-parquet-variantopt in explicitly via a sharedenable_all_registered_array_encodingstest helper.The
vortex/src/editions/tests.rsround-trip tests exercise every writer customization path (default strategy, replacement builders, opaque/probe compressors, custom flat strategies and field writers) to confirm none can emit an encoding outside the enabled editions.Testing
cargo nextest run -p vortex-file -p vortex-layout -p vortex -p vortex-parquet-variant— 410 passedcargo clippy --all-targets --all-features -p vortex-file -p vortex-layout -p vortex -p vortex-jni -p vortex-parquet-variant— cleancargo +nightly fmt --all🤖 Generated with Claude Code