Spring integration: jdbc-outbound-channel-adapter: how to ignore duplicatekey exception?

[H Paul]

	<int-jdbc:outbound-channel-adapter
		id="jdbcOutBoundAdapterEndpoint" channel="jdbc-out-channel"
		query="insert into MYTABLE (key, A, B) values(:payload[key], :payload[a],:payload[b])"
		data-source="outboundDataSource">
	</int-jdbc:outbound-channel-adapter>

When I have some duplicate rows (duplicate item/duplicate primary key), then I have below exception
and the whole process stopped. Meaning, no rows that are after the row which caused exception will be ignored.

ERROR LoggingHandler - org.springframework.integration.MessageHandlingException:
...
Caused by: org.springframework.dao.DuplicateKeyException:
...
Caused by: java.sql.SQLIntegrityConstraintViolationException:

Question:

Is there a way or what option should I use/configure the jdbc outbound-channel-adapter
to skip/ignore the DuplicateKeyException and continue the process.

Something like

	for loop {
		try 
		{
insert into MYTABLE ...
		} catch (DuplicateKeyException e) {
			//ignore it
		}
	}

[Jayesh A Lalwani]

All exceptions thrown by any channel goes to the Error handling channel. You can either configure a global error channel, or you can specify an error channel specific to a given channel. Look at this for explanation of error handling

[H Paul]

1. Read link. Very helpful. And tested below 2 cases. Still have question for CASE 2, (have a scan at CASE 1)

2. There are 2 cases:

CASE 1: "sender and reciever in the same thread" (SYNC)

	<!-- DirectChannel -->
	<int:channel id="jdbc-out-channel">
	</int:channel>

	<int-jdbc:outbound-channel-adapter
		id="jdbcOutBoundAdapterEndpoint" channel="jdbc-out-channel"
		query="insert into MYTABLE (key, A, B) values(:payload[key], :payload[A],:payload[B])"
		data-source="outboundDataSource">
	</int-jdbc:outbound-channel-adapter>

With the above config, I have a "sender and reciever in the same thread". So I can catch the exception and ignore it
and continue the next row insertion. This is OK but SYNCHRONOUS.

		listOfRows.stream()
		.map(rowMessage->MessageBuilder.withPayload(rowMessage).build())
		.peek(logger::debug)
		.forEach(message->{
				try {
					outchannel.send(message);
				} catch (Exception e) {
					//ignore it
					e.printStackTrace();  // Got what I wanted.... The sender of message got the exception thrown by receiver (jdbc outbound adapter end point)
				}
				}
			);

CASE 2: "sender and reciever are NOT in the same thread" (A-SYNC)

	<!-- QueueChannel -->
	<int:channel id="jdbc-out-channel">
		<int:queue/>
	</int:channel>

	<int-jdbc:outbound-channel-adapter
		id="jdbcOutBoundAdapterEndpoint" channel="jdbc-out-channel"
		query="insert into MYTABLE (key, A, B) values(:payload[key], :payload[A],:payload[B])"
		data-source="outboundDataSource">
				<int:poller fixed-rate="100000000"/>
	</int-jdbc:outbound-channel-adapter>	

With the above config, the sender send a message and go away. Meanwhile, when the receiver insert the rows
and if there is any kind of SQL/database exception related, it throw the exception and STOPPED there. Aborted the whole
process.

The exception is put into the 'errorChannel' either by default OR by custom. This is OK.

What I see is once the exception is put into the 'errorChannel', the whole process is stopped.

Question:
XML-wise: how I can tell Spring Integration, when there is an exception, ignore that row (Spring does via 'errorChannel')
and continue the next rows that follow the exception.