vortex-data/vortex - CodSpeed

vortex

Blog Docs Changelog

Performance History

Latest Results

Optimize list_sum empty/all-null fix-up, add list_sum bench Signed-off-by: Matt Katz <mhkatz97@gmail.com>

mk/list-sum-expr

2 hours ago

Implement compare in vortex instead of falling back to arrow (#8661) Instead of going through arrow compare logic implement it in vortex Lets us leverage optimised bitbuffer optimistaions and ultimately remove arrow dependency from vortex-array Signed-off-by: Robert Kruszewski <github@robertk.io>

develop

2 hours ago

Audit panic handling in detached tasks (#8677) ## Rationale for this change Following #8672, I've audited a few other suspicous code paths that had similar shape, mostly on the read side. ## What changes are included in this PR? 1. Propagate panics and shutdown correctly for `FileSegmentSource`. 2. Handle panics correctly in the shared iterator used by DuckDB 3. Misc panic handling around our runtime abstraction. ## What APIs are changed? Are there any user-facing changes? None --------- Signed-off-by: Adam Gutglick <adam@spiraldb.com>

develop

3 hours ago

spec refresh Documentation restructure (consolidated Specification section, developer guide grouped by audience, refreshed concepts/user-guide) plus a specification of the on-disk/on-wire encoding format (validity contract, per-encoding byte layouts, array/scalar/file/dtype framing) and a docs-conformance harness that derives the stable-encoding set from the code. Signed-off-by: mprammer <martin@spiraldb.com> Co-Authored-By: Claude <noreply@anthropic.com>

mp/spec

3 hours ago

ci: install only the uv binary for v4 ingest (drop the workspace sync) (#8517) The v4 Postgres ingest runs `uv run --no-project --with psycopg[binary] --with boto3 --with xxhash`, which needs only the uv binary and an ephemeral env — never the workspace. But the "Install uv" step uses `spiraldb/actions/setup-uv`, which runs a full `uv sync --all-extras --dev` that rebuilds `vortex-python` via maturin → sccache → S3. On the `extras=s3-cache` bench runners that fails `s3:GetObject AccessDenied` under the just-assumed ingest role (which only has `rds-db:connect`). It's `continue-on-error`, so CI never broke, and production confirmed the sync is dead weight — the ingest succeeded even when "Install uv" went red, because the binary installs before the sync and the ingest doesn't use the synced workspace. Switch the three v4 "Install uv" steps to `astral-sh/setup-uv` (the same v7.6.0 binary that action vendors) with no sync, which removes the wasteful workspace build and the sccache→S3 dependency entirely. The benchmark job's own non-v4 `setup-uv` is unchanged. Follow-up to #8516, which fixed the same failure by reordering so the sync ran with good credentials; this drops the now-pointless sync instead, so step order no longer matters. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Signed-off-by: Connor Tsui <connor@spiraldb.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

develop

3 hours ago

spec refresh Documentation restructure (consolidated Specification section, developer guide grouped by audience, refreshed concepts/user-guide) plus a specification of the on-disk/on-wire encoding format (validity contract, per-encoding byte layouts, array/scalar/file/dtype framing) and a docs-conformance harness that derives the stable-encoding set from the code. Signed-off-by: mprammer <martin@spiraldb.com> Co-Authored-By: Claude <noreply@anthropic.com>

mp/spec

3 hours ago

feat(geo): add WKT literal decoder; unify native WKB/WKT decoding Add native_geometry_scalar_from_wkt(&str), the WKT analogue of native_geometry_scalar_from_wkb: decode a single geometry literal to its native Point/LineString/Polygon/MultiPoint/MultiLineString/MultiPolygon scalar. Returns None for kinds without a native type (e.g. GeometryCollection) and Err for malformed input. Plan-time, one value only. Both decoders now detect the geometry kind from the parsed geoarrow element via geo_traits::GeometryTrait::as_type() and funnel through one shared dispatch and cast/import/wrap tail, rather than each format re-parsing the literal with its own parser. This reuses geo-traits (already a dependency) for kind detection, so no new crate is needed, and both WKB and WKT support all six OGC simple-feature types. Tests: an rstest over all six kinds for WKT, GeometryCollection -> None, malformed -> Err, and WKT == WKB for the same geometry. Signed-off-by: Nemo Yu <zyu379@wisc.edu>

nemo/geo-native-scalar-from-wkt

5 hours ago

ci: install only the uv binary for v4 ingest (drop the workspace sync) The v4 Postgres ingest runs `uv run --no-project --with ...`, which needs only the uv binary and an ephemeral env -- never the synced workspace. The "Install uv for v4 ingest" steps ran a full `uv sync --all-extras --dev` that rebuilt vortex-python via maturin -> sccache -> S3, forcing the fragile "ORDER MATTERS" constraint (sync before the ingest-role switch, whose creds lack S3). Since #8683 made the v4 ingest required, a spurious failure in that pointless build would fail the whole job. Pass `sync: false` to setup-uv in bench.yml and commit-metadata.yml (the established pattern elsewhere in .github), and delete the step outright in sql-benchmarks.yml, whose bench job already installs the uv binary. The ordering constraint disappears with the sync. Signed-off-by: Connor Tsui <connor@spiraldb.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KSAiqyHg1zjeX5ymFc292v

ct/bench-v4-emitter-uv-binary

7 hours ago

Latest Branches

+33%

Add `list_sum` expression#8676

3 hours ago

85b8055

mk/list-sum-expr

+12%

docs: documentation rework + Vortex encoding-format specification#8690

3 hours ago

42cce13

mp/spec

+11%

feat(geo): decode WKT geometry literals to native scalars#8689

7 hours ago

5dc69c1

nemo/geo-native-scalar-from-wkt

© 2026 CodSpeed Technology

Home Terms Privacy Docs