Copilot commented on code in PR #4814:
URL: https://github.com/apache/arrow-adbc/pull/4814#discussion_r4180462428
##########
python/adbc_driver_manager/adbc_driver_manager/dbapi.py:
##########
@@ -1290,12 +1303,12 @@ def adbc_read_partition(self, partition: bytes) -> None:
"""
_requires_pyarrow()
self._clear()
- self._results = None
+ self._cancel = self._conn._conn.cancel
handle = _blocking_call(
- self._conn._conn.read_partition, (partition,), {},
self._stmt.cancel
+ self._conn._conn.read_partition, (partition,), {}, self._cancel
)
self._rowcount = -1
- self._results = _RowIterator(self._stmt, handle, self._backend)
+ self._results = _RowIterator(self._cancel, handle, self._backend)
Review Comment:
The added Flight SQL test only cancels while `adbc_read_partition()` itself
is blocked, so no `_RowIterator` is ever created. It would still pass if this
line incorrectly used statement cancellation, leaving the new cancellation
behavior during a blocking `fetchone()`/`fetch_arrow_table()` unprotected.
Please add a partition response that returns a stream before later reads block,
then verify `adbc_cancel()` interrupts that fetch through the connection.
--
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]
