Skip to content

[pull] master from cube-js:master#451

Merged
pull[bot] merged 2 commits into
code:masterfrom
cube-js:master
May 6, 2026
Merged

[pull] master from cube-js:master#451
pull[bot] merged 2 commits into
code:masterfrom
cube-js:master

Conversation

@pull
Copy link
Copy Markdown

@pull pull Bot commented May 6, 2026

See Commits and Changes for more details.


Created by pull[bot] (v2.0.0-alpha.4)

Can you help keep this open source service alive? πŸ’– Please sponsor : )

ovr added 2 commits May 6, 2026 13:44
Clients can opt in via `HttpQuery.response_format = Arrow` to receive results as a binary Arrow IPC stream wrapped in a new `HttpQueryResult` flatbuffer variant, instead of the legacy `HttpResultSet` where every
cell is stringified.

Default stays `Legacy`, so existing clients that don't set the flag keep getting `HttpResultSet` β€” no behavior change for them.
parse_date_str` previously walked a 7-format `chrono::NaiveDateTime::parse_from_str`
cascade. Cube's canonical wire shape is `%Y-%m-%dT%H:%M:%S%.3f` β€”
`YYYY-MM-DDTHH:MM:SS.fff`, exactly 23 bytes β€” and was the 5th attempt in
that cascade, so most rows reparsed the prefix 4 times before matching.

Benchmark on Apple M3 Max (`cargo bench --bench transform_response`,
10k rows, master vs. this commit):

| Shape (10k rows)          | Before    | After     | Ξ”      |
| ------------------------- | --------- | --------- | ------ |
| row      / cols=8  / td=1 | 22.48 ms  | 15.84 ms  | -29.5% |
| columnar / cols=8  / td=1 | 7.76 ms   | 3.08 ms   | -60.3% |
| row      / cols=16 / td=2 | ~43 ms    | 36.19 ms  | -15.8% |
| columnar / cols=16 / td=2 | ~16 ms    | 6.11 ms   | -61.8% |
| columnar / cols=16 / td=0 | unchanged | unchanged | noise  |
@pull pull Bot locked and limited conversation to collaborators May 6, 2026
@pull pull Bot merged commit 7f37faa into code:master May 6, 2026
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant