Recordeset per message with variable recordset sequence?

Question

Hi,
I'm uploading a text-file with three different recordset substructures in any number and in variable sequence. I.e. a file could look like below:
structure 1
structure 1
structure 3
structure 2
structure 1
structure 3
structure 1
Thus the File adapter settings are as follows:
Recordset Structure = structure 1,,structure 2,,structure 3.*
Recordset Sequence = Variable
This works fine, the problem is the file is huge and I would like to limit the number of records in each message by using the recordset per message parameter. But if you have a variable recordset sequence the whole message is regarded as one recordset. If I change to ascending sequence I get an unpredictable number of messages since a recordset change is only done every time the file "breaks the order" defined in the Recordset Structure. Basically I want each record to be regarded as a recordset so I can split the file up with a predictable number of record per message. Is there any work around possible?
//Johan

Answer 1

Dear Johan,
if we say structure 1 as "A" structure 2 as "B" structure 3 as "C" and with sequence as variable then "regular expressions" reprsented  as
(ABC).
any combination at file structure comes as a message or as you say single record at sender side.
if you really want to break the records and conquer where order is not important,
that is A record ,B record, C record are independent of each other at XI processing  then go for 3 FILE ADAPTERS and 3 different message structures as A*  2)B* 3)C*.
this will definitely result in controlling message load at XI .
but can say for sure that there can be no replacement for (ABC) since it results in various combinations of A,B,C for example take computer language set it is (10)* and your is (ABC) there by its very diffciult to break this structure equivalent to some structure which meets your requirement.
and also  no meaning of breaking it as (AB)* or any combination,
we will end no where if we look for replacement for (ABC),
what can we do is go for independent structures or limit Structure order at FILE mean
atleast (ABC)* or (ABC) .
Thanks & Regards,
Rama Krishna