apache/druid · error · UnsupportedOperationException

Map column doesn't support getRow()

Error message

Map column doesn't support getRow()

What it means

MapTypeMapVirtualColumnDimensionSelector wraps a map-typed virtual column and only supports map-style access (getKeySelector/getObject). The DimensionSelector contract's row-oriented and dictionary-lookup methods are meaningless for a map column, so this class deliberately throws UnsupportedOperationException instead of returning wrong data.

Source

Thrown at extensions-contrib/virtual-columns/src/main/java/org/apache/druid/segment/MapTypeMapVirtualColumnDimensionSelector.java:49

/**
 * {@link DimensionSelector} for {@link Map} type {@link MapVirtualColumn}. This dimensionSelector only supports
 * {@link #getObject()} currently.
 */
final class MapTypeMapVirtualColumnDimensionSelector extends MapVirtualColumnDimensionSelector
{
  MapTypeMapVirtualColumnDimensionSelector(
      DimensionSelector keySelector,
      DimensionSelector valueSelector
  )
  {
    super(keySelector, valueSelector);
  }

  @Override
  public IndexedInts getRow()
  {
    throw new UnsupportedOperationException("Map column doesn't support getRow()");
  }

  @Override
  public ValueMatcher makeValueMatcher(@Nullable String value)
  {
    return new ValueMatcher()
    {
      @Override
      public boolean matches(boolean includeUnknown)
      {
        // Map column doesn't match with any string
        return false;
      }

      @Override
      public void inspectRuntimeShape(RuntimeShapeInspector inspector)
      {

View on GitHub (pinned to 9b90983fd2)

Solutions

1. Do not use the map virtual column as a scalar/array dimension; use its keys/values via MAP_KEYS() or MAP_VALUES() functions or index into it instead.
2. Rewrite the query so the map column is only consumed through getObject()/map-aware operators (e.g. SELECT it as JSON, or extract a specific key).
3. If you own custom query code, detect map-typed columns and avoid calling getRow()/lookupName()/idLookup() on their selectors.
4. Contribute upstream support (return an IndexedInts of the key/value selector) if a legitimate row-wise use is needed.

Example fix

// before
DimValHolder v = selector.getRow(); // throws UnsupportedOperationException
// after
Object mapVal = selector.getObject(); // map-aware access
// or extract a key in SQL: SELECT MAP_EXTRACT('k', mapCol) AS k FROM ...

// Java: before using the selector row-wise
if (selector instanceof MapTypeMapVirtualColumnDimensionSelector) {
  // use map-aware path
  Object mapValue = ((BaseMapVirtualColumnDimensionSelector) selector).getObject();
} else {
  IndexedInts row = selector.getRow();
}

boolean supportsRowAccess(DimensionSelector s) {
  return !(s instanceof MapTypeMapVirtualColumnDimensionSelector);
}

try {
  IndexedInts row = selector.getRow();
  process(row);
} catch (UnsupportedOperationException e) {
  processMapAware(selector.getObject());
}

When it happens

Trigger: Calling getRow() on a dimension selector obtained from a map-type virtual column (e.g. using MAP_KEYS/MAP_VALUES-backed virtual columns) in any query path that expects an array/list dimension: group-by on the column directly, filter code that iterates IndexedInts, or any engine/operator that reads the selector row-wise instead of via getObject().

Common situations: Users try to group by or filter directly on a map virtual column expression; query planners or custom extensions assume every DimensionSelector supports getRow; older query engines or community code paths written before map-typed columns existed reach this selector.

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

• Map column doesn't support lookupName()
• Map column doesn't support idLookup()
• Reverse lookup not allowed.
• Serialization not supported here
• Avro + JQ not supported

AI-assisted analysis of apache/druid@9b90983fd2 (2026-09-07). Data as JSON: /api/errors/feffd782368742b4. Report an issue: GitHub.