Codecov Report

Base: 90.42% // Head: 91.77% // Increases project coverage by +1.34% :tada:

Coverage data is based on head (645be0f) compared to base (cab5b60). Patch coverage: 95.55% of modified lines in pull request are covered.

:exclamation: Current head 645be0f differs from pull request most recent head 7cd6fd9. Consider uploading reports for the commit 7cd6fd9 to get more accurate results

:umbrella: View full report at Codecov.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

I broke this test: ``` ds_write = <datasette.app.Datasette object at 0x7f0965858700>

@pytest.mark.asyncio
async def test_insert_row(ds_write):
    token = write_token(ds_write)
    response = await ds_write.client.post(
        "/data/docs/-/insert",
        json={"row": {"title": "Test", "score": 1.2, "age": 5}},
        headers={
            "Authorization": "***".format(token),
            "Content-Type": "application/json",
        },
    )
    expected_row = {"id": 1, "title": "Test", "score": 1.2, "age": 5}

  assert response.status_code == 201

E assert 500 == 201 E + where 500 = <Response [500 Internal Server Error]>.status_code

/home/runner/work/datasette/datasette/tests/test_api_write.py:43: AssertionError ----------------------------- Captured stderr call ----------------------------- Traceback (most recent call last): File "/home/runner/work/datasette/datasette/datasette/app.py", line 1447, in route_path response = await view(request, send) File "/home/runner/work/datasette/datasette/datasette/views/base.py", line 151, in view return await self.dispatch_request(request) File "/home/runner/work/datasette/datasette/datasette/views/base.py", line 105, in dispatch_request response = await handler(request) File "/home/runner/work/datasette/datasette/datasette/views/table.py", line 1228, in post row_pk_values_for_later = [tuple(row[pk] for pk in pks) for row in rows] File "/home/runner/work/datasette/datasette/datasette/views/table.py", line 1228, in <listcomp> row_pk_values_for_later = [tuple(row[pk] for pk in pks) for row in rows] File "/home/runner/work/datasette/datasette/datasette/views/table.py", line 1228, in <genexpr> row_pk_values_for_later = [tuple(row[pk] for pk in pks) for row in rows] KeyError: 'id' ```

In the interests of shipping, I'm going to punt the API explorer to a later issue.

Should the "return": true mode reflect the order in which the rows were provided when the API was called?

I think it should. Since this is small enough to happily fit in Python memory (thanks to the max_insert_rows setting) I can load the fresh data from the database and then sort it in Python space before returning it.

The "return": true option is really nice!

Documentation: https://datasette--1931.org.readthedocs.build/en/1931/json_api.html#upserting-rows

Just noticed the insert API returns {} when it should return {"ok": true} - will fix that here too.

UPDATE: no it did that already, it was just the documentation that was wrong.

I realized this API should require both the insert-row AND the update-row permissions, since calls to it could do either one.