zhangrenhua commented on a change in pull request #32222:
URL: https://github.com/apache/spark/pull/32222#discussion_r625583986
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/jdbc/JDBCRDD.scala
##########
@@ -301,10 +306,44 @@ private[jdbc] class JDBCRDD(
ResultSet.TYPE_FORWARD_ONLY, ResultSet.CONCUR_READ_ONLY)
stmt.setFetchSize(options.fetchSize)
stmt.setQueryTimeout(options.queryTimeout)
- rs = stmt.executeQuery()
+
+ val isFinished = new AtomicBoolean(false)
+ startTheJobInterruptListenerThread(context, stmt, isFinished)
+ try{
+ rs = stmt.executeQuery()
+ } finally {
+ isFinished.set(true)
+ }
val rowsIterator = JdbcUtils.resultSetToSparkInternalRows(rs, schema,
inputMetrics)
CompletionIterator[InternalRow, Iterator[InternalRow]](
new InterruptibleIterator(context, rowsIterator), close())
}
+
+ /**
+ * Start the job interruption listener thread, if the job is interrupted and
+ * stmt.executeQuery() is not completed, initiate a stmt.cancel() request.
+ */
+ private def startTheJobInterruptListenerThread(context: TaskContext, stmt:
PreparedStatement
+ , isFinished: AtomicBoolean):
Unit = {
Review comment:
Try to write a complex SQL, such as Cartesian product to test the effect
of jdbc `fetchsize`，Or based on the jdbc query of hive on mapreduce, you will
find that the `fetchsize` is set at this time, and you have to wait for the job
to execute for a period of time before returning the cursor object. The client
cannot terminate the job run.
--
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.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
