prestodb/presto · error · RuntimeException

%s codec is not supported

Error message

%s codec is not supported

What it means

ParquetCompressor.getCompressor maps a Parquet CompressionCodecName to a ParquetCompressor implementation. Only GZIP, SNAPPY, ZSTD, and UNCOMPRESSED are wired up (LZO and LZ4 are explicitly unsupported TODOs). This RuntimeException is the fall-through guard thrown when the requested codec is anything else — i.e. you asked the writer to use a compression codec Presto's Parquet writer does not implement.

Source

Thrown at presto-parquet/src/main/java/com/facebook/presto/parquet/writer/ParquetCompressor.java:50

{
    ParquetDataOutput compress(BytesInput bytesInput)
            throws IOException;

    static ParquetCompressor getCompressor(CompressionCodecName codec)
    {
        // TODO Support LZO and LZ4 compression
        // When using airlift LZO or LZ4 compressor, decompressing page in reader throws exception.
        switch (codec.getParquetCompressionCodec()) {
            case GZIP:
                return new GzipCompressor();
            case SNAPPY:
                return new AirLiftCompressor(new SnappyCompressor());
            case ZSTD:
                return new AirLiftCompressor(new ZstdCompressor());
            case UNCOMPRESSED:
                return null;
        }
        throw new RuntimeException(String.format("%s codec is not supported", codec));
    }

    class GzipCompressor
            implements ParquetCompressor
    {
        @Override
        public ParquetDataOutput compress(BytesInput bytesInput)
                throws IOException
        {
            ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream();
            try (GZIPOutputStream outputStream = new GZIPOutputStream(byteArrayOutputStream)) {
                outputStream.write(bytesInput.toByteArray(), 0, toIntExact(bytesInput.size()));
            }
            byte[] bytes = byteArrayOutputStream.toByteArray();
            return createDataOutput(Slices.wrappedBuffer(bytes, 0, bytes.length));
        }
    }

View on GitHub (pinned to 55bb57d202)

Solutions

1. Switch the compression codec to a supported one: GZIP, SNAPPY, ZSTD, or UNCOMPRESSED (SNAPPY or ZSTD recommended)
2. Check the Presto config property controlling Parquet compression and correct the value
3. If LZ4/LZO is required, upgrade Presto or implement the codec in ParquetCompressor (see the TODO)
4. Wrap writer construction to fail fast with a clear config message instead of hitting the runtime fall-through

Example fix

// before
compressionCodec = CompressionCodecName.LZ4; // not implemented
// after
compressionCodec = CompressionCodecName.ZSTD; // supported by ParquetCompressor.getCompressor

// Java
import org.apache.parquet.hadoop.metadata.CompressionCodecName;
static boolean isSupportedParquetCodec(CompressionCodecName codec) {
    switch (codec) {
        case GZIP:
        case SNAPPY:
        case ZSTD:
        case UNCOMPRESSED:
            return true;
        default:
            return false;
    }
}

// Java
try {
    ParquetCompressor c = ParquetCompressor.getCompressor(codec);
} catch (RuntimeException e) {
    throw new IllegalArgumentException("Unsupported Parquet codec in config: " + codec, e);
}

When it happens

Trigger: Constructing a Parquet writer/compression setup with codec LZO, LZ4, BROTLI, or any other CompressionCodecName not in {GZIP, SNAPPY, ZSTD, UNCOMPRESSED}, causing getCompressor to fall through the switch.

Common situations: Configuring `hive.compression-codec` or writer properties to LZ4/BROTLI/LZO for Parquet output; copying a config from an ORC or other-format job where those codecs are valid; version drift where a codec enum value exists but the writer never implemented it.

Related errors

• DISTRIBUTED_TRACING_ERROR
• Applying decryptor on plaintext file
• Column encrypted with footer key: No keys available
• ACCUMULO_TABLE_EXISTS
• UNEXPECTED_ACCUMULO_ERROR

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/87669169d88fe3a5. Report an issue: GitHub.