Complex Flat File with Occurs in the table

karthegx · Post by **karthegx** » Thu Jan 24, 2013 2:27 am

Hi Guys,

I unable to find a way to derive this in the complex flat file . Below is the copy book

01 C-MSG-DTA-PORT-MRK.
10 C-MSG-DTA-PORT-MRK-TBL OCCURS 10 TIMES
INDEXED BY PORT-SUB.
15 C-SHP-ORD-CTL-NO PIC X(9).
15 C-PORT-MRK-LN-NO PIC 9(3).
15 C-PORT-MRK-TXT PIC X(40).
05 FILLER PIC X(274).

How can we achieve this ? Thanks in Advance

Kartheek

ray.wurlod · Post by **ray.wurlod** » Thu Jan 24, 2013 7:08 pm

Where is PORT-SUB defined?

Import the table definition from the FD.
Load the imported table definition into the Complex Flat File stage.
You can specify whether you want the OCCURS handled by creating a vector within a single column or by creating ten separate columns (flattening).

FranklinE · Post by **FranklinE** » Fri Jan 25, 2013 11:35 am

ray.wurlod wrote:Where is PORT-SUB defined?

Ray, in Cobol the index for OCCURS clauses is defined simply by using the clause. It is managed (as it were) by the compiler, and my sieve-like memory says that it only appears in the Assembled code as a register.

I'm not completely sure, but I believe it is similar to 88-level items.

It has a dedicated key-word, SET, to manipulate its value.

ray.wurlod · Post by **ray.wurlod** » Fri Jan 25, 2013 4:03 pm

Most of the ones I've seen have the value as a second-level (level 10 in this example) item.

FranklinE · Post by **FranklinE** » Fri Jan 25, 2013 4:13 pm

My COBOL/2 reference reminds me that indexing was a later innovation, but it didn't "retire" the original subscript usage. If the code contains an explicit variable for referencing the occurance, it is labelled a "subscript" and can be manipulated like any numeric field. I've seen code where the subscript was a better choice, since it might be used in more than one place.

It's like mixed metaphors. We can never be sure what the original intentions were.

ray.wurlod · Post by **ray.wurlod** » Fri Jan 25, 2013 7:47 pm

Sorry, I was thinking of OCCURS DEPENDING ON.

ray.wurlod · Post by **ray.wurlod** » Fri Jan 25, 2013 7:49 pm

So the index value is not really relevant - it's documentation only as far as a DataStage developer is concerned.

If you bring in the multi-valued field as a vector, then the vector will have an integer index of some kind automatically in DataStage, though the first element of the vector is position 0.

If you bring in the multi-valued field as separate columns, there will be ten of these.