Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-27589 Spark file source V2
  3. SPARK-27326

Fall back all v2 file sources in `InsertIntoTable` to V1 FileFormat

    XMLWordPrintableJSON

Details

    • Sub-task
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • 3.0.0
    • 3.0.0
    • SQL
    • None

    Description

      In the first PR for file source V2, there was a rule for falling back Orc V2 table to OrcFileFormat: https://github.com/apache/spark/pull/23383/files#diff-57e8244b6964e4f84345357a188421d5R34

      As we are migrating more file sources to data source V2, we should make the rule more generic. This PR proposes to:
      1. Rename the rule `FallbackOrcDataSourceV2 ` to `FallBackFileSourceV2`.The name is more generic. And we use "fall back" as verb, instead of "fallback".
      2. Rename the method `fallBackFileFormat` in `FileDataSourceV2` to `fallbackFileFormat`. Here we should use "fallback" as noun, while "fall back" is verb.
      3. Add new method `fallbackFileFormat` in `FileTable`. This is for falling back to V1 in rule `FallbackOrcDataSourceV2 `.

      Attachments

        Issue Links

          Activity

            People

              Gengliang.Wang Gengliang Wang
              Gengliang.Wang Gengliang Wang
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: