ponichWebKnight opened a new pull request, #40907:
URL: https://github.com/apache/superset/pull/40907
### SUMMARY
Adds header-click **server-side** sorting to the paginated/truncated table
views so
sorting reflects the **full dataset** rather than only the rows already
loaded on the
client. Previously, clicking a column header in these views either did
nothing useful or
only reordered the current page.
**What changed**
- **"View as table" results pane** (`DataTablesPane` → `GridTable`):
clicking a column
header now injects an `orderby` into the chart-data request and
re-queries the server.
- The server round-trip is **skipped** when the result already fits
within the row
limit (client-side sort is exact) — so small results stay instant.
- It is also **skipped for queries with post-processing** (pivot / cum /
rolling).
`orderby` + `row_limit` apply to the raw SQL *before* post-processing,
which would
change the rows feeding those operations and produce incorrect values
(e.g. a
cumulative total computed over a truncated subset). Those views fall
back to
client-side sorting of what the chart actually produced.
- The results pane stays mounted during a re-sort refetch, so the grid
keeps its sort
state/indicator instead of resetting.
- **Multi-column sort**: the chart-data API now validates only that each
`orderby`
direction is a boolean and allows multiple sort columns; the table
plugins'
(the metric-based default is only applied when no `orderby` is provided).
- **Drill to detail** (`DrillDetailPane`): column-header sort is wired
through to
`/datasource/samples` as a server-side `orderby`.
- `SamplesPayloadSchema` accepts an `orderby` field.
- `_get_drill_detail` preserves a caller-supplied `orderby`, defaulting
to the first
column only when none is given (previously it unconditionally forced
the first
column, ignoring the request).
- `orderby` is omitted from the `COUNT(*)` query (ordering a bare count
by a data
column is meaningless and errors on stricter databases).
**Design decisions / rationale**
- Server-side sort is only enabled where the displayed columns map
directly to the SQL
result; post-processed charts are intentionally left to client-side sort
to avoid
returning incorrect data.
- Sort-column validation is kept minimal at the API layer; unknown sort
columns are
rejected deeper in the query builder (`get_sqla_query` raises on
unresolved columns),
so injection-style column names cannot reach SQL.
### BEFORE/AFTER SCREENSHOTS OR ANIMATED GIF
<img width="868" height="541" alt="Screenshot 2026-06-09 at 18 14 20"
src="https://github.com/user-attachments/assets/dd912699-5818-4582-9a29-fd6c99b8e261";
/>
<img width="910" height="430" alt="Screenshot 2026-06-09 at 18 09 34"
src="https://github.com/user-attachments/assets/9ec51bcb-89cd-4913-9051-2ada26cd72aa";
/>
<img width="909" height="465" alt="Screenshot 2026-06-09 at 18 09 41"
src="https://github.com/user-attachments/assets/f278cf8a-80ac-43fc-8449-b0933932425d";
/>
<img width="874" height="482" alt="Screenshot 2026-06-09 at 18 09 59"
src="https://github.com/user-attachments/assets/a21fa5ee-4173-421a-a838-07b449d2dc3f";
/>
### TESTING INSTRUCTIONS
Automated:
- Backend unit tests:
- `pytest tests/unit_tests/charts/data/test_api.py`
- `pytest tests/unit_tests/views/datasource/utils_test.py`
- `pytest tests/unit_tests/common/test_query_actions.py`
- Backend integration test:
- `pytest tests/integration_tests/datasource_tests.py -k
samples_with_orderby`
Manual:
1. **Results pane (raw/non-post-processed chart):** open a chart whose
result exceeds the
row limit, open the **"View as table"** menu, click a column header →
confirm the
request re-fires with `queries[0].orderby` and the rows re-order by the
full dataset
(top-N changes), cycling asc → desc → off.
2. **Results pane (post-processed chart, e.g. Big Number with a cumulative
trend):** click
a column header → confirm sorting happens client-side with **no**
server request and
the displayed values stay correct (no truncated cumulative totals).
3. **Drill to detail:** right-click a chart → *Drill to detail*, click a
column header →
confirm the `/datasource/samples` request includes `orderby`, rows sort
server-side,
and sorting persists while paging.
4. **Multi-column sort:** Shift-click a second header in the results table
→ confirm
multiple `orderby` entries are sent and accepted.
### ADDITIONAL INFORMATION
<!--- Check any relevant boxes with "x" -->
<!--- HINT: Include "Fixes #nnn" if you are fixing an existing issue -->
- [ ] Has associated issue:
- [ ] Required feature flags:
- [x] Changes UI
- [ ] Includes DB Migration (follow approval process in
[SIP-59](https://github.com/apache/superset/issues/13351))
- [ ] Migration is atomic, supports rollback & is backwards-compatible
- [ ] Confirm DB migration upgrade and downgrade tested
- [ ] Runtime estimates and downtime expectations provided
- [x] Introduces new feature or API
- [ ] Removes existing feature or API
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
