apache/druid · error · java.lang.UnsupportedOperationException

FixedBucketsHistogramBufferAggregator does not support getLo

Error message

FixedBucketsHistogramBufferAggregator does not support getLong()

What it means

FixedBucketsHistogramBufferAggregator.getLong() unconditionally throws UnsupportedOperationException. A fixed-buckets histogram cannot be represented as a long, so any request to read the aggregation buffer as a long is invalid; the library fails fast rather than returning a meaningless value.

Source

Thrown at extensions-core/histogram/src/main/java/org/apache/druid/query/aggregation/histogram/FixedBucketsHistogramBufferAggregator.java:78

    innerAggregator.aggregate(buf, position, val);
  }

  @Override
  public Object get(ByteBuffer buf, int position)
  {
    return innerAggregator.get(buf, position);
  }

  @Override
  public float getFloat(ByteBuffer buf, int position)
  {
    throw new UnsupportedOperationException("FixedBucketsHistogramBufferAggregator does not support getFloat()");
  }

  @Override
  public long getLong(ByteBuffer buf, int position)
  {
    throw new UnsupportedOperationException("FixedBucketsHistogramBufferAggregator does not support getLong()");
  }

  @Override
  public double getDouble(ByteBuffer buf, int position)
  {
    throw new UnsupportedOperationException("FixedBucketsHistogramBufferAggregator does not support getDouble()");
  }

  @Override
  public void close()
  {
    // no resources to cleanup
  }

  @Override
  public void inspectRuntimeShape(RuntimeShapeInspector inspector)
  {
    inspector.visit("selector", selector);

View on GitHub (pinned to 9b90983fd2)

Solutions

Access the histogram via histogram-specific post-aggregators (quantile, min, max, sum) instead of raw long reads.
Ensure the query's output type for this metric is the complex/finalized histogram type, not LONG.
Refactor downstream code to consume FixedBucketsHistogram objects or their serialized base64 form.

Example fix

// before
Long v = agg.getLong(buf, pos);
// after: apply a scalar-extracting post-aggregator first
// {"type": "max", "fieldName": "myHist"}

Defensive patterns

Strategy: type-guard

Validate before calling

if (aggregator instanceof FixedBucketsHistogramBufferAggregator) {
  throw new IllegalStateException("Use histogram post-aggregators, not long reads");
}

Type guard

boolean supportsLong(BufferAggregator a) {
  return !(a instanceof FixedBucketsHistogramBufferAggregator);
}

Try / catch

try {
  value = aggregator.getLong(buf, position);
} catch (UnsupportedOperationException e) {
  Object v = aggregator.get(buf, position);
}

Prevention

Avoid projecting fixedBucketsHistogram metrics as BIGINT/LONG in SQL or native queries.
Use histogram post-aggregators (min, max, quantile, sum) to obtain numeric values.
Check the aggregator's comparator/type before numeric buffer access.

When it happens

Trigger: A query execution path or custom aggregation framework invokes getLong() on the buffer for a fixedBucketsHistogram metric, e.g. a long-typed expectation from the column signature or a numeric post-aggregator applied directly to the histogram metric.

Common situations: Pointing a LongSumPostAggregator or SQL BIGINT-typed projection at a fixedBucketsHistogram column; framework code assuming all aggregators produce numeric buffers.

Related errors

AI-assisted analysis of apache/druid@9b90983fd2 (2026-09-07). Data as JSON: /api/errors/8620c3e4c3104d26. Report an issue: GitHub.