Skip to content

Add support for an *input* format variants registry #209

Description

@Aklakan
  • For CSV variants we support the CSVW model
  • For optimal RDF/Trig processing we need to know e.g. whether all quads are represented as triples wrapped in a curly braces block - i.e. is it sufficient to only search for the pattern graph <g> { <s> <p> <o> } (regions always start near a curly brace) or do we need to consider <s> <p> <o> <g> . too.

What is needed is a way to conveniently configure format variants from the CLI.
Partly this is already possible via spark.hadoop.mapred.{input-format-prefix} settings that get picked up in the InputFormat implementation. But this is very cumbersome to configure. So an extra set of cli options and a registry is needed.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions