transformation using tMap etc. for positional file input.

One Star

transformation using tMap etc. for positional file input.

Hi All,
I am new to talenD and this forum so please excuse anomalies and/or gaffes in etiquette and provide guidance as necessary.
Also I started this thread because I could not find one specific to my question.
My talenD version is
Version: 3.1.2
Build id: r25815-20090619-1911

Problem statement - I have a positional file(fixed length) which has multiple record format types identified by group code '01', '02', '05' and '07'. I have to read these records and apply some tranformation logic and get only one output record format.
Such as :-
 if input record type is '01' then input record date is picked up and applied to output record or else current date is applied to output record.

 if input record type is '02' then read the 'clearing firm code' and copy to record type '05' and '07'.

 After applying above 2 transformations, get rid of record type '01' and '02'.

Here is what I think should be the design of solution should be :-
1. Use file positional metadata components to define different file format types.
2. Use tFilterRow component in 4 separate jobs/subjobs to read the same file and filter records format type '01','02','05','07'.I am not sure if this will result in parallel read of file. I need to check the java code.
2. Now since record type '02' needs to be read to apply logic to types '05' and '07', use tReplicate to replicate record type '02' and then use tMap for applying transformation to get the output record.
I am not sure if this is the cleanest design. I will post the images later.
Here is what I had already tried and failed :-
1. Take one input and filter it for record type '01', '02', '05' and '07'. Then use tReplicate to generate 4 outputs, then filter again to extract data for each separate record type '01', '02','05' and '07'.
However I stumbled on the next step where all these components do not take multiple inputs generated from same input source - tMap,tUnite,tNormalize.
Let me know what you guys think. Your input is appreciated.