Hi
Try with Camel 2.6 as we have fixed some issues with the splitter.
On Sun, Jan 23, 2011 at 2:23 PM, Christian Mueller
<christian.muel...@gmail.com> wrote:
>
> Hello list,
> I'm using Camel 2.4 (inside FUSE ESB 4.3). For a POC (prove of concept) I
> have to read a file with multiple address lines like this one:
> {code}
> 1,Hahnstr,1,60525,Frankfurt
> ...
> 10000,Hahnstr,10000,60525,Frankfurt
> {code}
>
> Afterwards I have to call a web service (which makes the address updates on
> our customer database). The web service respond with the id (the first
> number in the input file) and the status (UPDATED). I have to write a result
> file for each processed record which should looks like (the order of the
> entries is not relevant):
> {code}
> 1,UPDATED
> ...
> 10000,UPDATED
> {code}
>
> My first solution which works fine was the following:
>
> {code}
> from("file:src/test/data?noop=true").routeId("POC-main")
> .split(body().tokenize("\n"), new
> PocResultAggregationStrategy()).parallelProcessing().streaming()
> .inOut("direct:addressService")
> .end()
> .inOnly(targetFileResourcePath);
>
> from("direct:addressUpdateService").routeId("POC-addressUpdateService")
> .convertBodyTo(UpdateAddress.class)
> .inOut("cxf:bean:addressService")
> .convertBodyTo(String.class);
> {code}
>
> After thinking about this solution, it has a few cons:
> 1) I didn't use a persistent aggregator which could be result in loosing all
> aggregated messages if someone shutdown the ESB.
> 2) The Aggregator looks a bit 'overingenieurd' for this requirement because
> I only append new lines.
>
> I tryied to implement a smarter solution and came up with this one:
>
> {code}
> from("file:src/test/data?noop=true").routeId("POC-main")
> .split(body().tokenize("\n")).parallelProcessing().streaming()
> .inOut("direct:addressService")
> .inOnly("file:target/outbox?fileExist=Append")
> .end();
>
> from("direct:addressUpdateService").routeId("POC-addressUpdateService")
> .convertBodyTo(UpdateAddress.class)
> .inOut("cxf:bean:addressService")
> .convertBodyTo(String.class);
> {code}
>
> But unfortunately, the result file only contained 9897 lines instead of
> 10000. :-(
> I couldn't find an option in the file-component to lock the file, so that
> only one thread can manipulate the file in one time.
>
> My question is: Does the Camel file producer has an option for that? If not,
> should the file producer has an option like: writeLock=fileLock (similar to
> the readLock option of the file consumer)? Then maybe we also need an option
> to specify how long Camel should wait for the lock until it fails...
>
> What you are thinking?
>
> Thanks in advance for your feedback,
> Christian
>
> --
> View this message in context:
> http://camel.465427.n5.nabble.com/Using-Splitter-and-file-producer-with-option-fileExist-Append-failed-tp3353618p3353618.html
> Sent from the Camel - Users mailing list archive at Nabble.com.
>
--
Claus Ibsen
-----------------
FuseSource
Email: cib...@fusesource.com
Web: http://fusesource.com
Twitter: davsclaus
Blog: http://davsclaus.blogspot.com/
Author of Camel in Action: http://www.manning.com/ibsen/
