This is an automated email from the ASF dual-hosted git repository.
morningman pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris-website.git
The following commit(s) were added to refs/heads/master by this push:
new 60c549695c8 [community] weekly report 2026-07-20 (#4014)
60c549695c8 is described below
commit 60c549695c8d5f4da6b2732351a333634238f308
Author: Mingyu Chen (Rayner) <[email protected]>
AuthorDate: Mon Jul 27 16:49:13 2026 +0800
[community] weekly report 2026-07-20 (#4014)
Co-authored-by: morningman <[email protected]>
---
src/pages/community-report/_reports/2026-07-20.mdx | 372 +++++++++++++++++++++
1 file changed, 372 insertions(+)
diff --git a/src/pages/community-report/_reports/2026-07-20.mdx
b/src/pages/community-report/_reports/2026-07-20.mdx
new file mode 100644
index 00000000000..b68c44ba4e1
--- /dev/null
+++ b/src/pages/community-report/_reports/2026-07-20.mdx
@@ -0,0 +1,372 @@
+{/* Auto-generated by community-radar `report export-mdx` — do not edit by
hand.
+ `report` is English-only structured data laid out by doris-website. */}
+
+export const label = "Jul 20 – Jul 26, 2026";
+export const week = "Week 30, 2026";
+export const stats = [
+ {
+ "num": "72",
+ "label": "Merged PRs"
+ },
+ {
+ "num": "36",
+ "label": "New issues"
+ },
+ {
+ "num": "45",
+ "label": "Contributors"
+ }
+];
+
+export const report = {
+ "summary": {
+ "lead": "This week the Doris community merged 72 PRs with 45 contributors,
with theme momentum continuing to tilt toward Real-time Data Warehouse (65) and
Kernel & Engineering (58), followed by Multi-modal Lakehouse and Cloud-Native,
signaling a focused vertical push centered on the query engine and lakehouse
write path. Highlights concentrate on Iceberg/Paimon write maturity, expanded
CDC ingestion, and deeper cloud storage and async caching, while the Nereids
optimizer advances mul [...]
+ "highlights": [
+ {
+ "title": "Iceberg read/write path maturity: nested schema evolution,
Parquet VARIANT, and lakehouse write support",
+ "narrative": "Three upgrades harden the Iceberg path: nested-column
schema changes, native reading of Parquet VARIANT, and unified JNI/WAL routing
on the FileScannerV2 path. Doris gains stronger compatibility with Iceberg v3
semi-structured data and complex schemas, signaling a more mature lakehouse
write/read foundation.",
+ "prs": [
+ {
+ "num": 65329,
+ "title": "[feature](iceberg) Support nested column schema change",
+ "url": "https://github.com/apache/doris/pull/65329";
+ },
+ {
+ "num": 63192,
+ "title": "[feature](iceberg) Support reading Iceberg variant from
Parquet",
+ "url": "https://github.com/apache/doris/pull/63192";
+ },
+ {
+ "num": 66008,
+ "title": "[Feature](file scanner) Support JNI and WAL in V2",
+ "url": "https://github.com/apache/doris/pull/66008";
+ }
+ ]
+ },
+ {
+ "title": "Lakehouse write expansion to Paimon via JNI",
+ "narrative": "Paimon write support lands via a JNI bridge to the Java
SDK, covering INSERT INTO and INSERT OVERWRITE. The single-writer design with
delegated partition/bucket routing mirrors the Iceberg path, extending Doris's
multi-modal lakehouse write surface.",
+ "prs": [
+ {
+ "num": 65381,
+ "title": "[feature](paimon) Paimon write support via JNI",
+ "url": "https://github.com/apache/doris/pull/65381";
+ }
+ ]
+ },
+ {
+ "title": "CDC streaming jobs: OceanBase source and cloud-aware row
binlog with locality",
+ "narrative": "Streaming jobs now accept OceanBase CDC sources, while
row binlog gains an independent tablet model that preserves locality with base
tablets across cloud and hybrid clusters. CDC ingest broadens, and binlog
reliability and consistency improve in cloud deployments.",
+ "prs": [
+ {
+ "num": 65588,
+ "title": "[feature](streaming-job) Support OceanBase CDC streaming
jobs",
+ "url": "https://github.com/apache/doris/pull/65588";
+ },
+ {
+ "num": 65978,
+ "title": "[Feature](binlog): keep row binlog tablet locality
during scheduling",
+ "url": "https://github.com/apache/doris/issues/65978";
+ },
+ {
+ "num": 65810,
+ "title": "[feat](row binlog)：support cloud binlog and independent
binlog tablet",
+ "url": "https://github.com/apache/doris/pull/65810";
+ }
+ ]
+ },
+ {
+ "title": "Cloud storage depth: native OSS SDK, GCS/MinIO/Ozone
plugins, keyless GCS vaults, async file cache",
+ "narrative": "The cloud storage stack deepens with a native Alibaba
OSS C++ SDK, dedicated GCS/MinIO/Ozone filesystem plugins, keyless GCS vaults
via GKE Workload Identity, and async file cache writes on remote reads.
Object-store coverage and cold-query throughput both improve.",
+ "prs": [
+ {
+ "num": 61329,
+ "title": "[feature] (cloud) Add Alibaba Cloud OSS C++ SDK stack -
Part 1",
+ "url": "https://github.com/apache/doris/pull/61329";
+ },
+ {
+ "num": 65834,
+ "title": "[feat](fs) Add GCS, MinIO and Ozone filesystem plugins
on a shared S3 base",
+ "url": "https://github.com/apache/doris/pull/65834";
+ },
+ {
+ "num": 65433,
+ "title": "[feat](vault) Support keyless GCS vaults with GKE
Workload Identity",
+ "url": "https://github.com/apache/doris/pull/65433";
+ },
+ {
+ "num": 65658,
+ "title": "[feature](be) Add asynchronous file cache writes",
+ "url": "https://github.com/apache/doris/pull/65658";
+ }
+ ]
+ },
+ {
+ "title": "Nereids optimizer pushdowns: multi-DISTINCT CTE split,
statistics-driven eager distinct, lazy predicate materialization",
+ "narrative": "Three optimizer wins land: CTE splitting for
multi-DISTINCT aggregates, statistics-driven distinct pushdown into eager
aggregation, and statistics-aware lazy materialization of multi-stage
predicates. Plans for complex aggregations and multi-predicate scans get both
cheaper and faster.",
+ "prs": [
+ {
+ "num": 65664,
+ "title": "[improvement](nereids) Support CTE split for DISTINCT
aggregate functions",
+ "url": "https://github.com/apache/doris/pull/65664";
+ },
+ {
+ "num": 65099,
+ "title": "[improvement](eager-agg) Push down distinct aggregation
using statistics",
+ "url": "https://github.com/apache/doris/pull/65099";
+ },
+ {
+ "num": 64891,
+ "title": "[improvement](be) Add statistics-aware multi-stage
predicate lazy materialization",
+ "url": "https://github.com/apache/doris/pull/64891";
+ }
+ ]
+ }
+ ],
+ "numbers": {
+ "mergedPrs": 72,
+ "newIssues": 36,
+ "contributors": 45
+ }
+ },
+ "repos": [
+ {
+ "repo": "apache/doris",
+ "scenarios": [
+ {
+ "name": "Multi-modal Lakehouse",
+ "mergedNarrative": "The multi-modal lakehouse deepened around
Iceberg and native Parquet. FileScannerV2 gains a V2-owned native Parquet path
with vectorized fixed-width decode, dictionary range filtering, lazy
complex-residual materialization, and typed dictionary optimization, making hot
scans markedly faster. Iceberg deletion-vector stability, write-metric config,
partial name mapping, and v3 row-lineage coverage all harden, Paimon",
+ "merged": [
+ {
+ "num": 63719,
+ "title": "[test](regression) Add Iceberg v3 row lineage
coverage",
+ "url": "https://github.com/apache/doris/pull/63719";
+ },
+ {
+ "num": 65674,
+ "title": "[improvement](file) Build native Parquet scan path for
FileScannerV2",
+ "url": "https://github.com/apache/doris/pull/65674";
+ },
+ {
+ "num": 65676,
+ "title": "[fix](iceberg) make iceberg deletion vector stable.",
+ "url": "https://github.com/apache/doris/pull/65676";
+ },
+ {
+ "num": 65759,
+ "title": "[fix](topn) Resolve topn lazy materialization column
indexes for aliases",
+ "url": "https://github.com/apache/doris/pull/65759";
+ },
+ {
+ "num": 65782,
+ "title": "[fix](iceberg) Honor disabled write metrics",
+ "url": "https://github.com/apache/doris/pull/65782";
+ }
+ ],
+ "inProgressNarrative": "Multi-modal lakehouse work is closing
fine-grained gaps across Paimon, Iceberg, Hive, and external catalogs: Paimon
DDL/external table properties, reused Paimon JNI scanners to cut memory,
Iceberg partition-evolution schema fidelity, and HiveText handling of empty
arrays/maps. New-this-week patches add information_schema pruning for external
tables, lazy Parquet dictionary loading, and broader Arrow",
+ "inProgress": []
+ },
+ {
+ "name": "Real-time Data Warehouse",
+ "mergedNarrative": "Real-time DWH work tightened correctness and
concurrency safety: equivalent multi-column COUNT DISTINCT is normalized,
auto-partition results are guarded against concurrent DROP, async group commit
is backpressured by per-table WAL count, CDC end offset stays consistent with
progress, snapshot splits isolate per-table schemas, and the TOCTOU NPE in
partition pruning is",
+ "merged": [
+ {
+ "num": 65206,
+ "title": "[fix](agg) Normalize equivalent multi-column distinct
counts",
+ "url": "https://github.com/apache/doris/pull/65206";
+ },
+ {
+ "num": 65282,
+ "title": "[fix](fe) Guard auto partition result against
concurrent drops",
+ "url": "https://github.com/apache/doris/pull/65282";
+ },
+ {
+ "num": 65362,
+ "title": "[fix](be) Backpressure async group commit by table WAL
count",
+ "url": "https://github.com/apache/doris/pull/65362";
+ },
+ {
+ "num": 65415,
+ "title": "# [fix](topn lazy materialization) Disable topn lazy
materialization on non-light-schema-change tables",
+ "url": "https://github.com/apache/doris/pull/65415";
+ },
+ {
+ "num": 65472,
+ "title": "[fix](rewrite rule) Reject mismatched compare plan in
PullUpJoinFromUnionAll",
+ "url": "https://github.com/apache/doris/pull/65472";
+ }
+ ],
+ "inProgressNarrative": "Real-time DWH work newly landing this week
focuses on score queries, scan limits, plan/cache consistency, bucket-shuffled
unions, and NDV/distinct optimization: parallel rowid fetch for TopN,
table-stream change-type predicate pushdown, runtime-filter partition pruning,
and snapshot consistency. Together they prevent query hangs, TopN miscounts,
and stale runtime filters that matter most for",
+ "inProgress": []
+ },
+ {
+ "name": "Compute-Storage Separation & Cloud-Native",
+ "mergedNarrative": "The storage-compute separation stack gained a
wave of stability hardening: the recycler S3 client now sets a sane request
timeout to avoid curl-28 on slow DeleteObjects, file cache LRU restore uses
parser cursors plus two-phase initialization to remove O(N²) work, rebuilt VCG
warm-up jobs are cancelled on DROP, empty-rowset recycling",
+ "merged": [
+ {
+ "num": 64758,
+ "title": "[fix](cloud) set recycler S3 client requestTimeoutMs
to avoid curl-28 on slow DeleteObjects (mirror #49315)",
+ "url": "https://github.com/apache/doris/pull/64758";
+ },
+ {
+ "num": 65174,
+ "title": "[fix](file-cache): speed up LRU restore startup path",
+ "url": "https://github.com/apache/doris/pull/65174";
+ },
+ {
+ "num": 65426,
+ "title": "[fix](fe) Cancel rebuilt VCG warm up jobs on drop",
+ "url": "https://github.com/apache/doris/pull/65426";
+ },
+ {
+ "num": 65427,
+ "title": "[fix](filecache) fix flaky be UT for LRU dump",
+ "url": "https://github.com/apache/doris/pull/65427";
+ },
+ {
+ "num": 65571,
+ "title": "[fix](load) support compute_group in FE stream load
routing and planning",
+ "url": "https://github.com/apache/doris/pull/65571";
+ }
+ ],
+ "inProgressNarrative": "Cloud-native work is broadening the
separation layer: the Alibaba OSS native C++ SDK stack, CPU-aware S3 rate
limiting and dynamic core counts, the Cloud Table Stream control plane, and
single-rowset grouped compaction are all moving toward landing. New this week
brings S3 rate-limit regression coverage, cumulative-point advancement under
parallel cumulative",
+ "inProgress": []
+ },
+ {
+ "name": "Agent Observability",
+ "mergedNarrative": "Agent-oriented observability is being unified
under consistent FE semantics. This week `json_extract_no_quotes` is registered
as a true alias of `json_extract_string` so the planner shares one binding and
rewrite path, and VARIANT rows in distributed ordered TopN with lazy
materialization now surface sparse sub-columns before any root array, restoring
correct, explainable",
+ "merged": [
+ {
+ "num": 65380,
+ "title": "[fix](fe) Treat json_extract_no_quotes as
json_extract_string alias",
+ "url": "https://github.com/apache/doris/pull/65380";
+ },
+ {
+ "num": 65660,
+ "title": "[fix](variant) Prefer sparse variant fields over root
value",
+ "url": "https://github.com/apache/doris/pull/65660";
+ }
+ ],
+ "inProgressNarrative": "Agent observability is hardening the
inverted-index stack end-to-end: blocking legacy V1 format, landing a Kuromoji
Japanese analyzer (key), and introducing the SNII storage format (highlight) to
replace an aging layout. New this week, fixes around BM25 scoring, variant
semi-structured rows, sub-reader NULL bitmaps, and a stabilized score-cache
regression test are",
+ "inProgress": [
+ {
+ "num": 64667,
+ "title": "[feature](inverted-index) Add Japanese (Kuromoji)
morphological analyzer",
+ "url": "https://github.com/apache/doris/pull/64667";
+ },
+ {
+ "num": 66052,
+ "title": "[feature](inverted index) Add SNII inverted index
storage format",
+ "url": "https://github.com/apache/doris/pull/66052";,
+ "isNew": true
+ },
+ {
+ "num": 64071,
+ "title": "[feature](routine-load) support fill_missing_columns
for json format to auto-complete unspecified columns",
+ "url": "https://github.com/apache/doris/pull/64071";
+ }
+ ]
+ },
+ {
+ "name": "Ecosystem Integration",
+ "inProgressNarrative": "Ecosystem integration is making the
surrounding tooling more dependable: the dbt-doris adapter's broken test
dependencies blocking local development are fixed, and a new this week patch
corrects multi-table CDC Chinese garbling on non-UTF-8 client JVMs, so
cross-platform ingestion behaves consistently.",
+ "inProgress": [
+ {
+ "num": 64929,
+ "title": "[fix](dbt-doris) Fix broken test dependencies
preventing adapter development",
+ "url": "https://github.com/apache/doris/pull/64929";
+ },
+ {
+ "num": 65852,
+ "title": "[fix](cdc) Encode multi-table CDC records as UTF-8",
+ "url": "https://github.com/apache/doris/pull/65852";,
+ "isNew": true
+ }
+ ]
+ },
+ {
+ "name": "Security & Governance",
+ "inProgressNarrative": "Security & governance sees one new this
week: tighter masking of sensitive information in FE logs, extending existing
audit-SQL encryption and property-map masking to close a remaining channel
where credentials or keys could leak into logs.",
+ "inProgress": [
+ {
+ "num": 65914,
+ "title": "[fix](fe) Tighten sensitive info masking in FE logs",
+ "url": "https://github.com/apache/doris/pull/65914";,
+ "isNew": true
+ }
+ ]
+ },
+ {
+ "name": "Kernel & Engineering",
+ "mergedNarrative": "Kernel engineering advanced on three fronts: a
shared `RowKeyEncoder` deduplicates primary-key encoding across segment
writers, PDEP/AVX2 vectorize bit unpacking and the AVG serialized-state merge
becomes vectorized, and correctness fixes land for pre-epoch Arrow timestamps,
const-column block merge, BlockingQueue waiter accounting, array_difference
OOB, and brpc secondary-package alias ownership. CI converges on",
+ "merged": [
+ {
+ "num": 65492,
+ "title": "[refactor](storage) extract RowKeyEncoder: dedup key
encoding across both segment writers",
+ "url": "https://github.com/apache/doris/pull/65492";
+ },
+ {
+ "num": 65564,
+ "title": "[improvement](regression) Reduce Kerberos test
environment memory",
+ "url": "https://github.com/apache/doris/pull/65564";
+ },
+ {
+ "num": 65689,
+ "title": "[fix](be) Correct Arrow timestamps before Unix epoch
in DATETIMEV2 deserialization",
+ "url": "https://github.com/apache/doris/pull/65689";
+ }
+ ],
+ "inProgressNarrative": "Kernel engineering is especially busy this
week, headlined by tenant-level Colocation (key) to break multi-tenant replica
redundancy and the move of Parquet microbenchmarks into the performance CI gate
(key). Newly merged fixes span Arrow timezones, rowset write failures, JNI
exception poisoning across pooled scanner threads, external-table metadata
caches, and a",
+ "inProgress": [
+ {
+ "num": 64167,
+ "title": "[feat](colocate) support tenant-level colocation",
+ "url": "https://github.com/apache/doris/pull/64167";
+ },
+ {
+ "num": 65936,
+ "title": "[improvement](build) Run Parquet microbenchmarks in
performance CI",
+ "url": "https://github.com/apache/doris/pull/65936";,
+ "isNew": true
+ }
+ ]
+ }
+ ],
+ "demand": [
+ {
+ "name": "Real-time Data Warehouse",
+ "narrative": "The community emphasizes that row-binlog tablets must
remain co-located with their base tablets on the same backend and storage path
during scheduling, preventing replica repair or balancing from misaligning
upstream and downstream. This safeguards CDC pipeline consistency and
end-to-end low latency for real-time workloads.",
+ "refs": [
+ {
+ "num": 65978,
+ "title": "[Feature](binlog): keep row binlog tablet locality
during scheduling",
+ "url": "https://github.com/apache/doris/issues/65978";
+ }
+ ]
+ },
+ {
+ "name": "Ecosystem Integration",
+ "narrative": "The community wants dbt-doris to expose async
materialized views as a first-class dbt materialization, so users can manage
Doris's incrementally refreshed views through familiar dbt workflows. This
bridges dbt's transformation modeling with Doris's high-performance
precomputation, strengthening Doris's interoperability within modern ELT
stacks.",
+ "refs": [
+ {
+ "num": 65967,
+ "title": "[Feature] (dbt-doris) Support async materialized views
as a dbt materialization",
+ "url": "https://github.com/apache/doris/issues/65967";
+ }
+ ]
+ },
+ {
+ "name": "Kernel & Engineering",
+ "narrative": "The community requests a SHOW TSO STATUS SQL command
so clients can inspect the global TSO's physical and logical clocks without
HTTP calls. This improves observability, debugging, and operational automation,
delivering a meaningful usability upgrade to Doris's core engine.",
+ "refs": [
+ {
+ "num": 65849,
+ "title": "[Feature] Add SHOW TSO STATUS command",
+ "url": "https://github.com/apache/doris/issues/65849";
+ }
+ ]
+ }
+ ]
+ }
+ ]
+};
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
