Feature/ADBC Server#2

Open

borodark wants to merge 105 commits intomasterfrom

feature/arrow-ipc-api

Commits on Jan 13, 2026

Phase I
borodark
committed
Phase I
borodark
committed
Phase II
borodark
committed
Phase III
borodark
committed
Phase III
borodark
committed
python client works
borodark
committed
rust lint
borodark
committed
example rename
borodark
committed
GC
borodark
committed
GC
borodark
committed
GC
borodark
committed
examples
borodark
committed
clippy fixes
borodark
committed
real example
borodark
committed
WIP - before rebase
borodark
committed
some milestone
borodark
committed
solid Alpha perhaps
borodark
committed
masters one
borodark
committed
lint
borodark
committed
GC
borodark
committed
GC
borodark
committed
GC
borodark
committed
GC
borodark
committed
GC
borodark
committed
Fix clippy error: remove unused import in arrow_native server tests
borodark
committed
debug more
borodark
committed
actions
borodark
committed
fix(ci): Change unreferenced snapshots from reject to warn
borodark
committed
refactor(cubesql): Remove Cube server dependency from Arrow IPC tests
borodark
committed
feat(cubesql): Enable Arrow IPC integration tests
borodark
committed
clipptocracy
borodark
committed
Totalimento e mano
borodark
committed
ISSUE: shoud not hang up just for uncompilibe SQL
borodark
committed
fair point, clippy
borodark
committed
potential client SegFault fix
borodark
committed
Venture into pricise types
borodark
committed
Tradeoffs: going deeper into integers with adbc
borodark
committed
shrink and distill
borodark
committed
after rebase
borodark
committed
rework example build setup
borodark
committed
more clean rebuild
borodark
committed
feat(arrow-ipc): Add deep clean mode to rebuild script
borodark
committed
fix(arrow-ipc): Handle build dependencies in rebuild script
borodark
committed
fix(arrow-ipc): Add oclif manifest generation and skipLibCheck
borodark
committed
docs(arrow-ipc): Add build troubleshooting section to README
borodark
committed
fix(arrow-ipc): Use yarn tsc for proper TypeScript project builds
borodark
committed
typos
borodark
committed
WIP
borodark
committed
pivot to streaming arrow from gateway.ts
borodark
committed
feat(cubesql): Add CubeStore direct connection prototype

borodark
and
claude
committed
4. ✓ Demonstrated pre-aggregation selection logic
borodark
committed
feat(cubesql): Add end-to-end pre-aggregation selection demo

borodark
and
claude
committed
Next for MVP is cubesqlplanner integration for pre-aggregation selection. Everything else in the hybrid approach is working
borodark
committed
MVP_COMPLETE
borodark
committed
The foundation is now in place for transparent pre-aggregation routing
borodark
committed
proceed with phase 2
borodark
committed
POC ADBC Client → Arrow IPC (4445) → cubesqld → Pre-agg Matching → CubeStore Direct
borodark
committed
WIP
borodark
committed
proceed with SQL rewrite
borodark
committed
test_sql_rewrite
borodark
committed
POC direct to CubeStore
borodark
committed
Consider connection pooling for even lower latency ;-)
borodark
committed
x18 max speed up
borodark
committed
feat(cubesql): Add query result caching for Arrow Native server

borodark
and
claude
committed
docs(cubesql): Add comprehensive cache implementation documentation

borodark
and
claude
committed
docs(arrow-ipc): Add comprehensive cache implementation reflection

borodark
and
claude
committed
docs(arrow-ipc): Add executive summary of cache implementation success

borodark
and
claude
committed
I'nt not doing no JS
borodark
committed
PR clenup
borodark
committed
PR clenup
borodark
committed
fix: Address formatting and clippy warnings

borodark
and
claude
committed
feat(examples): Add Python performance tests for Arrow IPC cache

borodark
and
claude
committed
refactor(tests): Focus Python tests on CubeSQL vs REST HTTP API comparison

borodark
and
claude
committed
feat(tests): Add full materialization timing to Python performance tests

borodark
and
claude
committed
docs(arrow-ipc): Add comprehensive documentation and local verification setup

borodark
and
claude
committed
docs: Refocus from Cache to Arrow Native Server with optional cache

borodark
and
claude
committed
docs: Fix messaging - this PR introduces Arrow Native server, not enhances it
borodark
committed
docs: Remove PostgreSQL compatibility from new features list
borodark
committed
docs: Fix architecture diagrams - show port 4445 as NEW, not 4444
borodark
committed
arrow_native_client.py
borodark
committed
use CUBESQL_ARROW_RESULTS_CACHE_ terminology
borodark
committed
docs: Update next-steps.md with completed tasks
borodark
committed
Rebased and integrated with power_of_3
borodark
committed
changing Cargo.lock
borodark
committed
fix(cubejs-backend-native): Add missing pre_aggregations parameter to MetaContext::new()

borodark
and
claude
committed
fake transport fix
borodark
committed
Major terminology change
borodark
committed
Terminology: ADBC(Arrow Native) instead of Arrow Native or Arrow IPC
borodark
committed
on the way to pre-commit hooks
borodark
committed
used in ADBC live tests
borodark
committed
shrink AI blabbering
borodark
committed
cleanup
borodark
committed
The Train of Thoughts archived to Library of Claudius. Plus some local env performance fixes
borodark
committed
More to the Claudius Personal Library
borodark
committed
More to the Claudius Personal Library
borodark
committed
cleanup
borodark
committed
network settles it
borodark
committed
md
borodark
committed
update dev.Dockerfile
borodark
committed
The Conteinerisation of the solution
borodark
committed
uniphy
borodark
committed
integrate into DOCS
borodark
committed
realistic load tests
borodark
committed
python rest vs adbc
borodark
committed
chore: trim arrow-ipc example documentation
borodark
committed