Annotation Interface ParquetConfig


Annotation for configuring Parquet serializer/parser on classes and methods.
See Also:
  • Optional Element Summary

    Optional Elements
    Modifier and Type
    Optional Element
    Description
    Add bean types column.
    Compression codec used when writing ("UNCOMPRESSED" or "GZIP", case-insensitive).
    Maximum allowed wire-declared element count (file/row-group row counts, column-chunk value counts) when parsing.
    Maximum allowed aggregate decompressed-byte total across every Parquet page read during a single parse.
    Maximum allowed size (in bytes) of the whole Parquet input buffered into memory before parsing.
    Maximum allowed wire-declared length (in bytes) for a single Parquet page (compressed or uncompressed) when parsing.
    Page size in bytes.
    int
    Rank for application order.
    Row group size in bytes.
  • Element Details

    • compressionCodec

      Compression codec used when writing ("UNCOMPRESSED" or "GZIP", case-insensitive).

      Any other value fails fast with an IllegalArgumentException naming the bad value and the valid write codecs. This includes unknown names (typos) and "SNAPPY", which the parser can decode but the serializer cannot write. The parser detects the codec from the file and ignores this setting.

      Default:
      ""
    • rowGroupSize

      Row group size in bytes.
      Default:
      ""
    • pageSize

      Page size in bytes.
      Default:
      ""
    • addBeanTypes

      Add bean types column.
      Default:
      ""
    • maxLength

      Maximum allowed wire-declared length (in bytes) for a single Parquet page (compressed or uncompressed) when parsing.

      Guards against malformed or adversarial input where a small file declares a huge per-page byte size. Default is "268435456" (256 MiB). A value of "0" or less disables the cap.

      Returns:
      The annotation value.
      Default:
      ""
    • maxCount

      Maximum allowed wire-declared element count (file/row-group row counts, column-chunk value counts) when parsing.

      Guards against malformed or adversarial input where a small footer declares a huge row or value count. Default is "10000000" (10 million). A value of "0" or less disables the cap.

      Returns:
      The annotation value.
      Default:
      ""
    • maxInputLength

      Maximum allowed size (in bytes) of the whole Parquet input buffered into memory before parsing.

      Guards against an oversized input being buffered unbounded into memory. Default is "268435456" (256 MiB). A value of "0" or less disables the cap.

      Returns:
      The annotation value.
      Default:
      ""
    • maxDecompressedBytes

      Maximum allowed aggregate decompressed-byte total across every Parquet page read during a single parse.

      Guards against a decompression bomb assembled from many pages that each individually pass the per-page maxLength() cap. Default is "67108864" (64 MiB). A value of "0" or less disables the cap.

      Returns:
      The annotation value.
      Default:
      ""
    • rank

      int rank
      Rank for application order.
      Default:
      0