Add a way to process EBCDIC data in-place without converting to Spark

[yruslan]

Background

With this use case we would like to use the Cobol Parser and decoder in order to

• split input files by records,
• process each record using a custom code
• the custom code would extract some fields of each record, process it, and write them back in the same EBCDIC format
• The processor could be just a library outside of Spark framework
• The solution should support all types of file types (F, V, VB, custom record extractor, etc)
• Options used with spark-cobol should be ideally reused in the file processor as well

It seems most of building blocks for building the solution is already in place.

Feature

Add a way to process EBCDIC data in-place without converting to Spark.

Example

--

Proposed Solution