I did a sample.
Jan
/**
* user@camel.a.o (2014-06-12):
*
* I have a CSV file with structured data in it:
*
* F1,xxx,,xxx,xxx
* F2,XXX,XXX
* F2,XXX,XXX
* F1,xxx,,xxx,xxx
* F2,XXX,XXX
* F2,XXX,XXX
* F2,XXX,XXX
*
* There is a logical connection between the rows starting
* with F1 and the subsequent F2 rows. So, in the example above,
* the first 3 rows would form a logical unit of work and the
* final 4 rows would form a logical unit of work.
*
* Can camel be configured to read in and process this data in
* LUW groups and fail the whole LUW should any one of the
* rows in the LUW be in error?
*
* Andy
*
* @author Jan
*
*/
public class LogicalUnitOfWorkTest extends CamelTestSupport {
/*
* Base idea is adding an id to each line and changing the
* id value every time when an "F1" occurred.
* Then you could use an aggregator to collect them.
*/
@Override
protected RouteBuilder createRouteBuilder() throws Exception {
return new RouteBuilder() {
@Override
public void configure() throws Exception {
from("file:src/test/resources?fileName=logical-unit-of-work-data.csv&move=.done")
.split(body(String.class).tokenize("\n"))
.process(new WorkIdGenerator())
.aggregate(
header(WorkIdGenerator.WORK_ID),
AggregationStrategies.bean(LineAggregationStrategy.class,
"append")
)
.completionTimeout(1000)
.to("mock:out")
//TODO try something with failing the whole
logical-unit-of-work.
;
}
};
}
@Test
public void check() throws InterruptedException {
// 3 Exchanges: line0, line1-3, line4-9
getMockEndpoint("mock:out").expectedMessageCount(3);
File dir = new File("src/test/resources");
File bak = new File(dir, ".done");
File data = new File(bak, "logical-unit-of-work-data.csv");
try {
FileUtils.moveFileToDirectory(data, dir, false);
} catch (IOException e) {
//ignore
}
// Wait so file-component could scan.
Thread.sleep(5000);
assertMockEndpointsSatisfied();
}
}
/**
* Previously a splits the csv-file into lines.
* Use this strategy to combine the data with same
* work-id.
*/
public class LineAggregationStrategy {
public String append(String s1, String s2) {
return s1 + "\n" + s2;
}
}
/**
* Store a 'work-id' on each exchange.
* Value of that id is changed every time if a new
* "logical unit of work" starts (means, starts with 'F1').
*/
final class WorkIdGenerator implements Processor {
public static final String WORK_ID = "WorkID";
String id = next();
public String next() {
id = UUID.randomUUID().toString();
return id;
}
@Override
public void process(Exchange exchange) throws Exception {
if (exchange.getIn().getBody(String.class).startsWith("F1,")) {
next();
}
exchange.getIn().setHeader(WORK_ID, id);
}
}
> -----Ursprüngliche Nachricht-----
> Von: kraythe . [mailto:kray...@gmail.com]
> Gesendet: Donnerstag, 12. Juni 2014 17:06
> An: Camel Users List
> Betreff: Re: Logical Units of Work
>
> You could use an aggregator to group the rows or use scatter gather
> (well gather since you are already scattered) EIP in order to correlate
> the rows.
> As for failing "in error" that rather depends on the error. If an
> exception is thrown that is not that hard. Gather all of the rows,
> split, process, aggregate. On the splitter share the unit of work and
> any exception in that unit of work will fail the exchange. The question
> is what happens after the exchange fails.
>
> *Robert Simmons Jr. MSc. - Lead Java Architect @ EA* *Author of:
> Hardcore Java (2003) and Maintainable Java (2012)*
> *LinkedIn: **http://www.linkedin.com/pub/robert-simmons/40/852/a39
> <http://www.linkedin.com/pub/robert-simmons/40/852/a39>*
>
>
> On Thu, Jun 12, 2014 at 9:42 AM, AndyBell
> <andy.b...@dstglobalsolutions.com>
> wrote:
>
> > Hi
> >
> > I've not used Camel for a number of years and now have a (to me)
> > non-trivial task which I wonder if it can be done in Camel using
> > standard camel
> > component:
> >
> > I have a CSV file with structured data in it:
> >
> > F1,xxx,,xxx,xxx
> > F2,XXX,XXX
> > F2,XXX,XXX
> > F1,xxx,,xxx,xxx
> > F2,XXX,XXX
> > F2,XXX,XXX
> > F2,XXX,XXX
> >
> > There is a logical connection between the rows starting with F1 and
> > the subsequent F2 rows. So, in the example above, the first 3 rows
> > would form a logical unit of work and the final 4 rows would form a
> > logical unit of work.
> >
> > Can camel be configured to read in and process this data in LUW
> groups
> > and fail the whole LUW should any one of the rows in the LUW be in
> error?
> >
> > Andy
> >
> >
> >
> > --
> > View this message in context:
> > http://camel.465427.n5.nabble.com/Logical-Units-of-Work-
> tp5752206.html
> > Sent from the Camel - Users mailing list archive at Nabble.com.
> >
