Flow, Document List to Document transformations...

[roberto.simoni]

Hi to all.
I'm trying to read a flat file and do some document transformation.
I'm reading the file supposing i'm goind to handle large files.
The problem is, when i have my raw document (from flat file) i try to
convert in a more convenient document where for example i "transform"
two fields (a date part and a time part) in a date field of java type set
as date.
To do this, i have trouble with document list and document transformations.
I have this steps:

== parseFile ===================================

pub.file:getFile
REPEAT
pub.flatFile:convertToValues
pub.flow:tracePipeline (*)
MyPackage.myFlow:transformRowToItem
pub.flow:tracePipeline (**)


== transformRowToItem ==========================

input .....: rawDocument
recordWithNoID(List)
ddate: String
dhour: String
output ....: itemDocument
thedate: Object[java.util.Date])

LOOP over '/row/recordWithNoID'
MAP (here handle transformation and map results to the output item)


What I would like to obtain is to have a unique itemDocument in the
pipeline at point (**) with a wonderful thedate field as java.util.Date

Instead, if I go in "trace here" mode till "transformRowToItem" all run
correctly. If I run all "parseFile" flow an error occur saying
"Dimension mismatch, List expected".

I'm using webmethods 6.5
I have tried some forum suggestions without success. I have tried to
use documentListToDocument but I don't know how to use it... absolutely
stacked.

Thanks in advance for any help.

[roberto.simoni]

Found the answer. The problem is the raw document type created
automatically from the flat file schema. That doc type has a recordWithNoID
set as DocumentList, but when you read file record by record in a REPEAT
step it is more better to drop that document type and create a new one
with recordWithNoID set up as Document.
Mapping ffValues of convertToValues to the new document type give a
unique raw document (as expected) and after that, you can add simply a
MAP step with transformations (so the LOOP step can be removed).
This is more readable and run correctly.
Hoping this could help someone else.
Bye