Package dev.frostlake.executor.copy

Interface StageFileReader

All Known Implementing Classes:
JsonStageReader, XmlStageReader
public interface StageFileReader
Reads a staged file of a record-based FILE_FORMAT (JSON, XML, Avro, Parquet, ORC) into a list of JsonNode records. The COPY loader turns each record into a row via its existing JSON row-building path, so the record formats share one loading path. CSV, being positional text rather than records, keeps its own path in the loader.

Readers are discovered at runtime as a ServiceLoader SPI: each implementation is a provider listed in META-INF/services/dev.frostlake.executor.copy.StageFileReader and is keyed by formatType(). The zero-dependency JSON and XML readers ship in the engine; the heavier Avro/Parquet/ORC readers live in the optional frostlake-formats module and light up only when it is on the classpath — so the core engine carries no Hadoop/Avro/ORC dependencies. Adding a format is "implement this interface and register the provider".

  • Method Summary

    Modifier and Type
    Method
    Description
    String
    formatType()
    The FILE_FORMAT TYPE this reader handles (e.g.
    List<tools.jackson.databind.JsonNode>
    readRecords(Path file)
    Parse file into its records.

  • Method Details

    • readRecords

      List<tools.jackson.databind.JsonNode> readRecords(Path file) throws IOException
      Parse file into its records. An unreadable/malformed file throws (a bad load is never silent).
      Throws:
      IOException

    • formatType

      String formatType()
      The FILE_FORMAT TYPE this reader handles (e.g. "JSON"), matched case-insensitively.