Page History

...

...
from(...)
   .aggregate()
       .correlationExpression(header(id))
       .aggregationStrategy(myStrategy)
       .completionTimeout(10000)
       .messageStore(myStore)
...

Claim check

THe claim check pattern temporarily reduces the data volume of the message by storing content in a message store in exchange for a claim check token. The content is retrieved later on before it's needed again.

// Optionally: override default store from context
// (ohr:) IMHO I don't think that this configuration level is really necessary
defaultMessageStore(myStore);

// 1) Store body.
// 2) Set body to null.
// 3) Set Exchange.CLAIM_CHECK header to unique claim id.
from(...)
   .claimCheck()  // store body in default store
   // .claimCheck(header('bigHeader'), customStore) : store header in custom store 
   .to(...);

// 1) Lookup for the Exchange.CLAIM_CHECK header value.
// 2) Read the message.
// 3) Set body to the value fetched from the store.
from(...)
   // .setHeader(Exchange.CLAIM_CHECK, const("id")) : header should still contain the claim id 
   .claim() // read body from default store
   // .transform(claim()) : more generically as expression
   // .setHeader('bigHeader', claim(customStore)) : retrieve from custom store back into original header
   .to(...);

Open issues:

• exception handling if there's no data available for a specific token
• clean up of stale content that was never claimed back
• maybe return some kind of DataHandler instead of a token (cf. CXF MTOM attachments) and retrieve content transparently?