Hello,
I am trying to create MT940 swift messages using a common format schema. The requirement is to split swift messages as chunks of 25. (To simplify the case, lets say it is 2). My input file is a delimited flat file. So I need a way to split a common format message like.
Header
Detail BEGIN
Record
Record
Record
Detail END
Trailer
into two seperate swift files (with 2 records and 1 record) such as
New Header
Detail Begin
Record
Record
Detail End
New Trailer
and
New Header
Detail Begin
Record
Detail End
New Trailer
I have been searching for possible solutions and came up with 3 different ways to accomplish this:
1. Create a map to split the records (in the same file) and then split the file using envelope debatching. My first question: is creating such a map is possible with Biztalk mapper (custom XSL maybe? – since the input and output schemas are complex, I’d prefer Biztalk mapper solution)? Ie the output will be:
<FILES>
<FILE>
Contents of first file ( has 2 records)
</FILE>
<FILE>
Contents of first file (has 1 record)
</FILE>
</FILES>
2. A custom pipeline solution: http://www.codeproject.com/KB/biztalk/DebatchingFlatfile.aspx
3. Orchestration XPath Debatching: http://www.biztalkgurus.com/media/p/15506.aspx
For the scope of this project, I might be dealing with some large files. (500.000 records = 50 MB input, even more later on) so performance is really an issue here. The second and third options might be easier to implement as there are some samples around but using xpath, memorystreams, etc for such large files might be causing some performance problems or even out of memory exceptions.
Any suggestions are welcome..