Repository navigation
Docs for 26.11.1.post3: the four glue changes with their measured figures, and the stale statements (refs #315) - #399
Merged
Merged
Conversation
…, and the stale statements (refs #315) JVM iter_dicts parses a batch with no tagged value without the object hook (#382); JVM insert_arrow and insert_dataframe carry a timestamp or date column as one long[] of epoch ticks (#383); native transaction calls return the state as one integer (#384); native save() of a saved record with auto-transaction on skips the transaction check (#385). The figures are copied by script from the scored gate study and each is checked against the public comment on #315. The n_update_one figure ran with auto-transaction off, so the page says off (the comment says on). The save() paragraph said three calls without the auto-transaction case. The quickstart was re-run on both published 26.11.1.post3 packages on Python 3.12 and printed the output shown. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
tae898
marked this pull request as ready for review
November 14, 2026 07:14
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
I update the user docs for
26.11.1.post3, which has been on PyPI since 2026-11-14 for both packages. The four glue changes of the release are described where a reader looks for them, with the measured figures of the release gate: JVMiter_dicts()parses a batch with no tagged value without the object hook (#382), JVMinsert_arrowandinsert_dataframecarry a timestamp or date column as onelong[]of epoch ticks (#383), nativebegin,commit,rollback,with db.transaction()andis_transaction_active()get the transaction state as one integer (#384), and nativesave()of a saved record skips the transaction check with auto-transaction on (#385). The engine is unchanged.This PR is a draft because the pages describe code that is on the tag
26.11.1.post3and not yet onmain. It is ready once #387 (the merge of the post3 line intomain) is merged. The branch is based onmain, not on the tag, becausemainhas 24 doc pages and the tag has 16. The PR changes no code.The pages:
how-to/bulk-loads.adoc: how a timestamp or date column crosses on the JVM package, with the figures forinsert_arrowandinsert_dataframe, theinsert_columnsfigure that shows no claim, and the peak resident size, which was higher with the new path.how-to/databases-and-queries.adoc: theiter_dicts()change with its three figures, and the stalesave()sentence that said three calls without the auto-transaction case.how-to/transactions.adoc: a new section on what a transaction costs on the native build, with three figures.how-to/dataframes-and-columns.adoc: one sentence that points to the bulk-loads page for the difference between the packages.how-to/concurrency.adocandquickstart.adoc: the version that the pages name (26.11.1.post3).The repository has no changelog or release-notes page, so there is nothing to add for the earlier entries' format. The GitHub Release of each tag is generated by
release-notes.yml.What I verified
The behavior text comes from the diff of the tag
26.11.1.post3against26.11.1.post2(results.py,dataframe.py,core.py,TemporalColumn.java,DocumentBatcher.java,_native/database.py,_native/lib.py,_native/records.py,ArcadeNative.java) and from the tests of those changes.Every figure was copied by script from the scored gate study and checked against the text of my comment of 2026-11-14 on #315, so each ratio on the pages is one that the issue shows. The pages state the conditions: per operation, one Linux x86_64 laptop with two cores, Python 3.12, ABBA order, 20 clean pairs, an A/A control, and the wheels that CI built from the tagged commit. The published wheels were checked for the same engine as
26.11.1and were not timed again. The pages claim nothing forto_list()(its figures did not resolve), nothing for memory, and nothing forsave()with auto-transaction on.One correction to my comment on #315: it labels the native
set()andsave()figure as auto-transaction on. The study script never callsset_auto_transaction, so that figure ran with auto-transaction off, and the transactions page says so.quickstart.adocran on Python 3.12 against the publishedarcadedb 26.11.1.post3andarcadedb-jvm 26.11.1.post3, and all three scripts printed the output on the page. The local Antora 3.2.1 build withdocs/antora-playbook.ymlexits 0 with no warning.uvx pre-commit run --fileson the six files passes; its Python, YAML and shell hooks have no file to check here.I did not change the kill probe sentence in
bulk-loads.adoc("It has not been run on26.11.1.post2"). It is still true, and I have no run of it on post3.Related issues
Refs #315
Performance changes
None in the code. The figures on the pages are the ones of the release gate in #315.
Docs
The pages listed above.
Checklist
uvx pre-commit run --files <changed files>passes🤖 Generated with Claude Code