SEPURI-SAI-KRISHNA commented on issue #12649:
URL: https://github.com/apache/seatunnel/issues/12649#issuecomment-6057655078
Thanks. The fix is up as #12650 and now carries the coverage you specified.
The multi-batch case asserts values in order rather than the row count,
which matters because correcting only the loop bound leaves the vector indexed
past its end and arrow answers that with null instead of an exception. Measured
on the committed test, the bound-only version gives `expected: <[10, 11, 20,
null, 22]> but was: <[10, 11, 22, null, null]>`, so a row-count assertion alone
would have passed it.
Both column types you named are in: an `ARRAY<INT>` `ListVector` with
varying list lengths, and a `TimeStampMicroVector`, each written across the
batch boundary. The plan is `{2, 0, 3}`, so the empty middle batch covers the
skip branch of `readArrow()` that must not advance the read row count, and one
cell is deliberately null so a genuine null stays distinguishable from an
out-of-range read. `getReadRowCount()` is asserted too, since that is what both
BE readers use to advance their scan offset.
**On the Doris and StarRocks tests.** They pass. From the `test-reports`
artifact of run `37641679675` at the current head `d928cd54d`:
```text
connector-doris 15 classes 130 tests 0 failures 0 errors
connector-starrocks 13 classes 90 tests 0 failures 0 errors
```
Worth one note on how I got there, because my first attempt was misleading.
Locally those same suites reported 27 and 33 errors, which looked alarming
until I read them: every one was `Mockito cannot mock this class ... Java :
21`, with zero assertion-style failures. That is this machine's JDK, not the
reader. The CI counts above are the same 130 and 90 tests, all green on JDK 8,
which is the answer to your question rather than my local numbers.
For completeness: that `unit-test (8, ubuntu-latest)` leg does show red
overall, on `JobMetricsTest` in `seatunnel-engine` failing to join a Hazelcast
cluster. It is unrelated to this reader, which has only those two callers, and
I have triaged it on the PR.
Keeping this open until #12650 merges.
--
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]
