prestodb/presto · error · ParquetCorruptionException

Codec not supported in Parquet:

Error message

Codec not supported in Parquet: 

What it means

ParquetCompressionUtils.decompress dispatches on the Parquet column-chunk codec; codecs outside SNAPPY/GZIP/LZO/LZ4/ZSTD hit the default branch and throw ParquetCorruptionException 'Codec not supported in Parquet'. Either the enum has an unsupported member or the page header was misread.

Source

Thrown at presto-parquet/src/main/java/com/facebook/presto/parquet/ParquetCompressionUtils.java:66

        if (input.length() == 0) {
            return EMPTY_SLICE;
        }

        switch (codec) {
            case GZIP:
                return decompressGzip(input, uncompressedSize);
            case SNAPPY:
                return decompressSnappy(input, uncompressedSize);
            case UNCOMPRESSED:
                return input;
            case LZO:
                return decompressLZO(input, uncompressedSize);
            case LZ4:
                return decompressLz4(input, uncompressedSize);
            case ZSTD:
                return decompressZstd(input, uncompressedSize);
            default:
                throw new ParquetCorruptionException("Codec not supported in Parquet: " + codec);
        }
    }

    private static Slice decompressSnappy(Slice input, int uncompressedSize)
    {
        byte[] buffer = new byte[uncompressedSize];
        decompress(new SnappyDecompressor(), input, 0, input.length(), buffer, 0);
        return wrappedBuffer(buffer);
    }

    private static Slice decompressZstd(Slice input, int uncompressedSize)
    {
        byte[] buffer = new byte[uncompressedSize];
        decompress(new ZstdDecompressor(), input, 0, input.length(), buffer, 0);
        return wrappedBuffer(buffer);
    }

    private static Slice decompressGzip(Slice input, int uncompressedSize)

View on GitHub (pinned to 55bb57d202)

Solutions

1. Upgrade Presto to a build supporting the codec (e.g., LZ4_RAW, BROTLI support).
2. Rewrite/convert the Parquet file with SNAPPY or GZIP compression.
3. Recompress with a supported codec using parquet-tools/Spark before loading.
4. Verify file metadata isn't corrupted (re-transfer, validate with parquet-tools meta).

Example fix

// before: file written with BROTLI
spark.write.parquet(...).option("compression", "brotli")
// after
spark.write.parquet(...).option("compression", "snappy")

CompressionCodecName codec = footer.getFileMetaData().getCreatedBy() != null ? chunk.getCodec() : null;
EnumSet<CompressionCodecName> ok = EnumSet.of(SNAPPY, GZIP, LZO, LZ4, ZSTD);
if (!ok.contains(chunk.getCodec())) throw new IllegalArgumentException("unsupported codec: " + chunk.getCodec());

try { ParquetCompressionUtils.decompress(codec, input, size); } catch (ParquetCorruptionException e) {
    throw new DataCorruptionException("unsupported parquet codec", e);
}

When it happens

Trigger: A Parquet file's column chunk metadata specifies a codec this Presto build cannot decode (e.g., BROTLI or a newer codec like LZ4_RAW on old versions).

Common situations: Files written by tools using BROTLI or LZ4_RAW compression; old Presto version lacking newer codec support; corrupted footer metadata yielding a bogus enum.

Related errors

• NOT_SUPPORTED
• NOT_SUPPORTED
• NOT_SUPPORTED
• INVALID_TABLE_PROPERTY
• NOT_SUPPORTED

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/1a7c5460f04f6672. Report an issue: GitHub.