kerwin-zk opened a new pull request, #8527:
URL: https://github.com/apache/paimon/pull/8527
### Purpose
The Daft connector's `read_paimon` / `explain_paimon_scan` previously only
supported time travel by `snapshot_id` and `tag_name`. This adds a `timestamp`
argument so users can time-travel to the latest snapshot at or before a given
time — the third form Paimon already supports at the core level
(`scan.timestamp-millis` / `scan.timestamp`). Only the Daft wrapper was
missing; core is unchanged.
`timestamp` is dispatched by type:
- `int` (epoch milliseconds) or `datetime.datetime` ->
`scan.timestamp-millis` (a naive `datetime` is interpreted in the local
timezone)
- `str` (e.g. `'2026-07-09 10:00:00'`) -> `scan.timestamp`
`snapshot_id`, `tag_name` and `timestamp` are mutually exclusive; setting
more than one raises `ValueError`.
```python
read_paimon("db.orders", opts, timestamp="2026-07-09 10:00:00")
read_paimon("db.orders", opts, timestamp=datetime(2026, 7, 9, 10))
read_paimon("db.orders", opts, timestamp=1751990400000)
```
### Tests
Added to `tests/daft/daft_integration_test.py`:
- `test_read_paimon_with_timestamp` — end-to-end: with a cutoff between two
commits, both the millis and `datetime` forms return only the first snapshot.
- `test_timestamp_scan_option_mapping` — unit coverage of the type→option
mapping (int / str / datetime; `bool` and `float` rejected).
- `test_read_paimon_rejects_multiple_time_travel` — parametrized
mutual-exclusion checks (replaces the old two-way check).
All `tests/daft/` pass locally (163 passed, 1 skipped); flake8 clean under
`dev/cfg.ini`.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
