Docs: add DataFusion backend support column to PPL command reference - #5679
Merged
noCharger merged 3 commits intoAug 7, 2026
Merged
Conversation
noCharger
requested review from
LantaoJin,
RyanL1997,
Swiddis,
acarbonetto,
ahkcs,
anirudha,
dai-chen,
joshuali925,
mengweieric,
penghuo,
ps48,
qianheng-aws,
songkant-aws,
vamsimanohar,
ykmr1224 and
yuancu
as code owners
August 5, 2026 16:45
Contributor
PR Reviewer Guide 🔍(Review updated until commit 6d2bf9c)Here are some key observations to aid the review process:
|
Add a 'DataFusion Backend' column to the Commands table in docs/user/ppl/index.md, indicating whether each PPL command runs on the analytics engine (DataFusion) execution path used for parquet/composite indices (Yes / Partial / No / N/A), with a legend defining the values. Signed-off-by: Louis Chu <lingzhichu.clz@gmail.com>
ahkcs
reviewed
Aug 6, 2026
makeresults lowers to a Values node whose DataFusion pushdown (VirtualTableScan, esp. CHAR/VARCHAR string literals) is not yet supported on main -- that fix is tracked in the (unmerged) core PR opensearch-project/OpenSearch#22554. Mark No until it lands. Signed-off-by: Louis Chu <lingzhichu.clz@gmail.com>
noCharger
force-pushed
the
docs/ppl-datafusion-backend-support
branch
from
August 6, 2026 07:37
2f5a6e4 to
6d2bf9c
Compare
Contributor
|
Persistent review updated to latest commit 6d2bf9c |
Per review by @ahkcs on PR opensearch-project#5679: - search -> Partial (relevance/full-text predicates are Lucene-delegated; base scan runs on DataFusion) - convert -> Partial (some conversion functions are not wired on the DataFusion path) - flatten -> No (struct/object flattening is not supported on parquet/composite) - bin -> Partial (span bucketing runs; other bin variants are not supported) expand and mvexpand were already changed to No in a prior commit per the same review. Signed-off-by: Louis Chu <lingzhichu.clz@gmail.com>
ahkcs
approved these changes
Aug 7, 2026
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
2 participants
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.
Description
Adds a DataFusion Backend column to the Commands table in
docs/user/ppl/index.md, indicating whether each PPL command runs on the analytics engine (DataFusion) execution path used forparquet/composite-format indices, plus a legend defining the values.Column values:
Basis
Values were derived from (a) the analytics-engine integration-test suite, (b) a live run of each command against a
parquet/composite index on an analytics-engine cluster, and (c) lowered-plan (Calcite RelNode) analysis for commands without direct coverage.Notable classifications:
parse— regex named-group extraction currently returns empty on the DataFusion path, whilerexextraction works. (Partial)streamstats—percentile/percentile_approxwindow functions are rejected; other aggregates work. (Partial)patterns—mode=aggregationis not yet supported;mode=labelworks. (Partial)lookup— the join-based form is not yet supported; the table form works. (Partial)multisearch— 3-branch union under-aggregates; 2-branch works. (Partial)xyseries— pending direct verification. (Partial)expand/mvexpand— lower to UNNEST (Correlate + Uncollect); the DataFusion path is not yet verified end-to-end (no analytics-engine IT coverage, and a table-registration gap fails multi-shard). (No)foreach— not supported on the DataFusion path in the current build. (No)makeresults— lowers toValues; the DataFusion pushdown (VirtualTableScan, in particular CHAR/VARCHAR string literals) is not yet supported and is tracked in the unmerged core PR [analytics-backend-datafusion] Normalize CHAR/VARCHAR literals to Str in VirtualTable conversion OpenSearch#22554. (No)mvcombine/nomv— the backend lacksARRAY_AGG/ARRAY_COMPACT. (No)ml/kmeans/ad/graphlookup— no DataFusion plan translation / unsupported on the Calcite path. (No)explain/describe/show datasources/rest— planning, metadata, or management commands that do not touch the parquet/AE data path (they run via the default engine and are unaffected by AE). (N/A)Check List
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.