Reuseable Jobs with EBCDIC data

[gsherry1]

Suppose I wish to create a job that will take input filename for an EBCDIC source as parameter and convert this to ASCII format. I wish to call this job with many different source formats. In each case I wish the input file to transform the datatypes to most appropriate non-mainframe datatype. For example, I would like COMP-3 packed decimals decoded to regular displayable ascii numerics, EBCDIC string converted to ASCII string etc.

1. Can EBCDIC be converted to ascii within datastage without the CFF component? I don't want to use this component, since it doesn't seem support much parameterization (schema file etc).

2. Can conversion between EBCDIC and ASCII be done without explicitly writing a derivation for each mapping. Ie. If I provide an EBCDIC description in the source schema file, and an ASCII description in the target schema will DS automatically convert? I have used ETL tools in the past that have done the transformation implicitely if your source was described EBCDIC and your target ASCII.

3. Can question 2 above be done with RCP, and have the data convert without any mention of the fields in the transform?

Unfortunately, I do not have a Parallel environment to test this out at this time.

Thanks.

[vmcburney]

When you import the definition of a complex flat file into DataStage it derives a DataStage data type for each field using the file definition. For cobol files this is the cfd file. All your mainframe data types will be converted to DataStage data types. Quite an easy task since all fields are written as fixed width text within the flat file.

The complex flat file input stage has a drop down combo box on the Record Options table that you set to ASCII or EBCDIC. The stage will automatically convert the EBCDIC data into a DataStage format. You also have a Text/Binary combo option and a Little-endian/Big-endian/Native-endian combo option and a decimal rounding option.

So the conversion is almost entirely implicit.

[gsherry1]

Thanks vmcburney.

Problem with the CFF stage is it does not seem to support schema files, and thus makes it difficult to parameterize processing involving sources with variations in column formats.

The orchestrate shell guide 4.5 seems to suggest that describing EBCDIC sources is possible with a schema file. The Import and Modify sections of the orchestrate documentation seem to also suggest that implicit conversion from EBCDIC to ASCII might be possible.

I was wondering if anybody has done such conversion with schema files without using CFF stage.