prestodb/presto · error · PrestoException

PARQUET_UNSUPPORTED_ENCODING

PARQUET_UNSUPPORTED_ENCODING

Error message

Column: %s, Encoding: %s

What it means

PrestoException (PARQUET_UNSUPPORTED_ENCODING) thrown at the end of Decoders.createValuesDecoder() when no branch matched the page's (type, encoding) combination at all — the encoding is not supported by the batch reader for this column. Unlike the type-specific errors, this signals an entirely unhandled encoding (or a supported encoding on a type outside its allowed branches).

Source

Thrown at presto-parquet/src/main/java/com/facebook/presto/parquet/batchreader/decoders/Decoders.java:297

        if (encoding == DELTA_BYTE_ARRAY && type == PrimitiveTypeName.FIXED_LEN_BYTE_ARRAY) {
            if (isDecimalType(columnDescriptor)) {
                ByteBufferInputStream inputStream = ByteBufferInputStream.wrap(ByteBuffer.wrap(buffer, offset, length));
                ValuesReader parquetReader = getParquetReader(encoding, columnDescriptor, valueCount, inputStream);

                if (isShortDecimalType(columnDescriptor)) {
                    return new FixedLenByteArrayShortDecimalDeltaValueDecoder(parquetReader, columnDescriptor);
                }

                return new FixedLenByteArrayLongDecimalDeltaValueDecoder(parquetReader);
            }
            else if (isUuidType(columnDescriptor)) {
                ByteBufferInputStream inputStream = ByteBufferInputStream.wrap(ByteBuffer.wrap(buffer, offset, length));
                ValuesReader parquetReader = getParquetReader(encoding, columnDescriptor, valueCount, inputStream);
                return new FixedLenByteArrayUuidDeltaValuesDecoder(parquetReader);
            }
        }

        throw new PrestoException(PARQUET_UNSUPPORTED_ENCODING, format("Column: %s, Encoding: %s", columnDescriptor, encoding));
    }

    private static ValuesReader getParquetReader(ParquetEncoding encoding, ColumnDescriptor descriptor, int valueCount, ByteBufferInputStream inputStream)
            throws IOException
    {
        ValuesReader valuesReader = encoding.getValuesReader(descriptor, VALUES);
        valuesReader.initFromPage(valueCount, inputStream);
        return valuesReader;
    }

    private static FlatDecoders readFlatPageV1(DataPageV1 page, RichColumnDescriptor columnDescriptor, Dictionary dictionary)
            throws IOException
    {
        byte[] bytes = page.getSlice().getBytes();
        ByteBuffer byteBuffer = ByteBuffer.wrap(bytes, 0, bytes.length);
        FlatDefinitionLevelDecoder definitionLevelDecoder = createFlatDefinitionLevelDecoder(
                page.getDefinitionLevelEncoding(),
                columnDescriptor.isRequired(),

View on GitHub (pinned to 55bb57d202)

Solutions

1. Dump the file's encoding per column with parquet-tools to identify the unhandled encoding.
2. Upgrade Presto to a version supporting that encoding.
3. Rewrite the file with standard encodings (PLAIN/RLE/dictionary).
4. Fall back to the legacy (non-batch) Parquet reader if available.

Example fix

// before
// unknown encoding -> PARQUET_UNSUPPORTED_ENCODING
// after
// force supported encodings in writer config
 writer.withEncoding(PLAIN) // or PLAIN_DICTIONARY/RLE_DICTIONARY

// enumerate encodings in the file footer and check support before reading
Set<Encoding> used = pages.stream().map(p -> p.getEncoding()).collect(toSet());
used.removeAll(SUPPORTED_ENCODINGS); // PLAIN, RLE, PLAIN_DICTIONARY, RLE_DICTIONARY, DELTA_*
if (!used.isEmpty()) throw new IllegalArgumentException("Unsupported encodings: " + used);

try {
    decoder = Decoders.readFlatPage(page, columnDescriptor, dictionary);
} catch (PrestoException e) {
    if (e.getErrorCode().getName().equals("PARQUET_UNSUPPORTED_ENCODING")) {
        // rewrite the file with standard encodings or upgrade Presto
    }
    throw e;
}

When it happens

Trigger: createValuesDecoder() (via valuesDecoder, createValuesDecoderV1/V2 paths) receives a ParquetEncoding not covered by PLAIN, RLE-boolean, dictionary, DELTA_BINARY_PACKED, DELTA_BYTE_ARRAY/DELTA_LENGTH_BYTE_ARRAY, or the UUID delta branch — e.g., exotic encodings or new encodings from newer writers.

Common situations: Reading files produced with encodings introduced after your Presto version; writers using BYTE_STREAM_SPLIT or other niche encodings; encrypted/extended encodings.

Related errors

• HUDI_CANNOT_OPEN_SPLIT
• Unsupported encoding
• PARQUET_IO_READ_ERROR
• PARQUET_IO_READ_ERROR
• PARQUET_UNSUPPORTED_COLUMN_TYPE

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/aefff60301316818. Report an issue: GitHub.