apache/iceberg · error · ParquetDecodingException

could not read page in col " + desc + " as the dictionary wa

Error message

could not read page in col " + desc + " as the dictionary was missing for encoding " + dataEncoding

What it means

VectorizedPageIterator.initDataReader starts decoding a data page whose encoding uses a dictionary, but no dictionary has been set for this column chunk. Iceberg throws ParquetDecodingException because dictionary-encoded pages cannot be decoded without their dictionary page. This usually means the dictionary page was skipped, lost, or the file is malformed.

Source

Thrown at arrow/src/main/java/org/apache/iceberg/arrow/vectorized/parquet/VectorizedPageIterator.java:77

  private DictionaryDecodeMode dictionaryDecodeMode;

  public void setAllPagesDictEncoded(boolean allDictEncoded) {
    this.allPagesDictEncoded = allDictEncoded;
  }

  @Override
  protected void reset() {
    super.reset();
    this.valuesReader = null;
    this.vectorizedDefinitionLevelReader = null;
  }

  @Override
  protected void initDataReader(Encoding dataEncoding, ByteBufferInputStream in, int valueCount) {
    ValuesReader previousReader = (ValuesReader) valuesReader;
    if (dataEncoding.usesDictionary()) {
      if (dictionary == null) {
        throw new ParquetDecodingException(
            "could not read page in col "
                + desc
                + " as the dictionary was missing for encoding "
                + dataEncoding);
      }
      try {
        dictionaryEncodedValuesReader =
            new VectorizedDictionaryEncodedParquetValuesReader(
                desc.getMaxDefinitionLevel(), setArrowValidityVector);
        dictionaryEncodedValuesReader.initFromPage(valueCount, in);
        if (ParquetUtil.isIntType(desc.getPrimitiveType()) || !allPagesDictEncoded) {
          dictionaryDecodeMode = DictionaryDecodeMode.EAGER;
        } else {
          dictionaryDecodeMode = DictionaryDecodeMode.LAZY;
        }
      } catch (IOException e) {
        throw new ParquetDecodingException("could not read page in col " + desc, e);
      }

View on GitHub (pinned to 86d9c8fc54)

Solutions

Ensure the full column chunk including the dictionary page is read (don't skip pages when iterating).
Validate and rewrite the corrupt Parquet file if its dictionary page is genuinely missing.
Upgrade Iceberg if a known bug in dictionary page handling matches your version.
Disable vectorized reads as a fallback to the standard Parquet reader.

Defensive patterns

Strategy: try-catch

Try / catch

try {
  // vectorized read
} catch (ParquetDecodingException e) {
  if (e.getMessage().contains("dictionary was missing")) {
    // re-read the full column chunk including dictionary page, or fall back to non-vectorized reader
  } else throw e;
}

Prevention

Always read complete column chunks; never skip dictionary pages
Validate files for missing dictionary pages before commit
Upgrade Iceberg if your version has known dictionary-page handling bugs

When it happens

Trigger: initDataReader receives a dataEncoding.usesDictionary()==true while the dictionary field is null for the column.

Common situations: Corrupted files missing dictionary pages, custom readers that skip dictionary pages, or row-group/page iteration that started mid-chunk after the dictionary page.

Understand the failure class

Background: 'Could not be found', 'does not exist', 'not found in database': the resource-not-found family when an ID, slug, key, or URI lookup comes back empty — this error's family across 20 libraries.

Related errors

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/1021ce4a091a25e8. Report an issue: GitHub.