apache/druid · error · UnsupportedOperationException

DimensionSpec[ ] cannot vectorize

Error message

DimensionSpec[%s] cannot vectorize

What it means

DimensionSpec.decorate for vectorized selectors (SingleValueDimensionVectorSelector) throws UOE("DimensionSpec[%s] cannot vectorize") by default; only dimension specs that support vectorization override it. When the vectorized query engine processes a column with a non-vectorizable DimensionSpec, this error aborts the query.

Solutions

1. Disable vectorization for the query: set "vectorize": false (or vectorizeDimension:false) in the query context
2. Replace the DimensionSpec with a vectorization-compatible one (e.g. DefaultDimensionSpec) or pre-transform data at ingestion
3. Implement/override decorate(SingleValueDimensionVectorSelector) and decorate(MultiValueDimensionVectorSelector) in your custom DimensionSpec
4. Check server logs/druid.query.vectorize config to confirm which component refused to vectorize

Example fix

// before
{"queryType":"timeseries", "dimensions":[{"type":"myCustomSpec",...}], ...}
// after
{"queryType":"timeseries", "dimensions":[{"type":"default","dimension":"page",...}], "context":{"vectorize":false}}

// reject non-vectorizable specs up front
if (!spec.getType().equals("default") && query.getContextBoolean("vectorize", true)) {
  log.warn("DimensionSpec may not vectorize; set vectorize=false or use DefaultDimensionSpec");
}

boolean canVectorize(DimensionSpec s) { return s instanceof DefaultDimensionSpec || s instanceof IndexedDimensionSpec; }

try { return factory.makeVectorSelector(spec); } catch (UnsupportedOperationException | UOE e) { return factory.makeDimensionSelector(spec); /* non-vector path */ }

When it happens

Trigger: Running a vector-enabled query (default when possible) whose groupBy/timeseries/topN uses a DimensionSpec that lacks vector support — e.g. certain extraction-fn or decorated specs, or custom DimensionSpec implementations.

Common situations: Custom DimensionSpec plugins not updated for vectorized engines; query JSON using extraction functions incompatible with vectorization; cluster where vectorization auto-enabled but specs predate it.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors

• DimensionSpec[ ] cannot be vectorized
• Aggregator[ ] cannot vectorize
• Cannot handle datasource
• Cannot handle subquery
• Cannot operate on a dimension with no dictionary

AI-assisted analysis of apache/druid@9b90983fd2 (2026-09-07). Data as JSON: /api/errors/a3e6a4e906674918. Report an issue: GitHub.

Appendix: source

Thrown at processing/src/main/java/org/apache/druid/query/dimension/DimensionSpec.java:78

  /**
   * Decorate a {@link DimensionSelector}, allowing custom transformation of underlying behavior (e.g. performing
   * extraction functions in the case of {@link ExtractionDimensionSpec}, regex filtering in the case of
   * {@link RegexFilteredDimensionSpec}, and so on).
   */
  DimensionSelector decorate(DimensionSelector selector);

  /**
   * Vectorized analog of {@link #decorate(DimensionSelector)} for {@link SingleValueDimensionVectorSelector}, most
   * likely produced with
   * {@link org.apache.druid.segment.vector.VectorColumnSelectorFactory#makeSingleValueDimensionSelector(DimensionSpec)}
   *
   * Decoration allows a {@link DimensionSpec} to customize the behavior of the underlying selector, for example
   * transforming or filtering values.
   */
  default SingleValueDimensionVectorSelector decorate(SingleValueDimensionVectorSelector selector)
  {
    throw new UOE("DimensionSpec[%s] cannot vectorize", getClass().getName());
  }

  /**
   * Vectorized analog of {@link #decorate(DimensionSelector) for {@link MultiValueDimensionVectorSelector}, most likely
   * produced with
   * {@link org.apache.druid.segment.vector.VectorColumnSelectorFactory#makeMultiValueDimensionSelector(DimensionSpec)}
   *
   * Decoration allows a {@link DimensionSpec} to customize the behavior of the underlying selector, for example
   * transforming or filtering values.
   */
  default MultiValueDimensionVectorSelector decorate(MultiValueDimensionVectorSelector selector)
  {
    throw new UOE("DimensionSpec[%s] cannot vectorize", getClass().getName());
  }

  /**
   * Does this DimensionSpec require that decorate() be called to produce correct results?
   */

View on GitHub (pinned to 9b90983fd2)