Hi Sounds like a job for the splitter EIP instead. http://camel.apache.org/splitter
Then you can use a java method as the splitter expression. Then you can implement logic to put together beloning lines. On Thu, Sep 13, 2012 at 9:39 AM, <klaus.schroe...@ethalon.de> wrote: > Hello Camel experts > we are using Camel to read a delimited file and import the data in our > system. The file has a quite simple structure: > > ABF31|01|10503547153|1|9921|90|3|3|20120713|20120713|20120718|0|0| > ABF37|01|10503547153|1|9921|65|1|1|0||PSB031|20120718||1 > ABF37|01|10503547153|1|9921|75|2|2|0||PSB031|20120718||1 > ABF37|01|10503547153|1|9921|80|2|2|0||PSB031|20120718||1 > ABF31|01|10503547154|1|9921|90|3|3|20120730|20120730|20120815|0|0| > ABF37|01|10503547154|1|9921|65|1|1|0||PSB031|20120815||1 > ABF37|01|10503547154|1|9921|95|1|1|0||PSB031|20120815||1 > ABF37|01|10503547154|1|9921|100|3|3|0||PSB031|20120815||1 > ABF31|01|10503547155|1|9921|90|3|3|20120730|20120730|20120807|0|0| > ABF37|01|10503547155|1|9921|75|3|3|0||PSB031|20120807||1 > ... > > So we have a abf31 record that has abf37 records as "children". We are > reading the file using Camel CSV data format and then want to use an > aggregator to aggregate the rows. The example above should result in 3 > abf31 objects. > Unfortunately the records do not have a real correlation identifier to be > used in the aggregator. The aggregation is complete if the next record is > an abf30 record. In "normal" programming this kind of grouping is quite > simple, but we did not find an obvious and easy way to do this with Camel > and aggregators. > > As a workaround we now try to insert a correlation id while reading the > CSV. Is this the recommended way to do it? > > Any help is appreciated. > Klaus > > > -- Claus Ibsen ----------------- FuseSource Email: cib...@fusesource.com Web: http://fusesource.com Twitter: davsclaus, fusenews Blog: http://davsclaus.com Author of Camel in Action: http://www.manning.com/ibsen
