apache/druid · error · IllegalStateException

Cannot translate sqlTypeName[%s] to Druid type for field[%s]

Error message

Cannot translate sqlTypeName[%s] to Druid type for field[%s]

What it means

When planning an SQL aggregation over an HLL sketch, HllSketchBaseSqlAggregator translates the input column's RelDataType (SQL type) into a Druid ColumnType. If the SQL type has no Druid equivalent (Calcites.getColumnTypeForRelDataType returns null), an IllegalStateException is thrown naming the sqlTypeName and field. This happens for SQL types the sketch aggregator cannot consume as pre-aggregated input.

Source

Thrown at extensions-core/datasketches/src/main/java/org/apache/druid/query/aggregation/datasketches/hll/sql/HllSketchBaseSqlAggregator.java:140

                        .orElse(false)) {
      aggregatorFactory = new HllSketchMergeAggregatorFactory(
          aggregatorName,
          columnArg.getDirectColumn(),
          logK,
          tgtHllType,

          // For HllSketchMergeAggregatorFactory, stringEncoding is only advisory to aid in detection of mismatched
          // merges. It does not affect the results of the aggregator. At this point in the code, we do not know what
          // the input encoding of the original sketches was, so we set it to the default.
          HllSketchAggregatorFactory.DEFAULT_STRING_ENCODING,
          finalizeSketch || SketchQueryContext.isFinalizeOuterSketches(plannerContext),
          ROUND
      );
    } else {
      final RelDataType dataType = columnRexNode.getType();
      final ColumnType inputType = Calcites.getColumnTypeForRelDataType(dataType);
      if (inputType == null) {
        throw new ISE(
            "Cannot translate sqlTypeName[%s] to Druid type for field[%s]",
            dataType.getSqlTypeName(),
            aggregatorName
        );
      }

      final DimensionSpec dimensionSpec;

      if (columnArg.isDirectColumnAccess()) {
        dimensionSpec = columnArg.getSimpleExtraction().toDimensionSpec(null, inputType);
      } else {
        String virtualColumnName = virtualColumnRegistry.getOrCreateVirtualColumnForExpression(
            columnArg,
            dataType
        );
        dimensionSpec = new DefaultDimensionSpec(virtualColumnName, null, inputType);
      }

View on GitHub (pinned to 9b90983fd2)

Solutions

Ensure the sketch input column is of a supported Druid type (string base64 sketch or numeric) and re-index segments with wrong types
Use HLLSketchBuild/other SQL helpers with the correct argument (e.g. DC or COUNT(DISTINCT) on a supported column)
If a custom type, convert/cast the expression to a supported type in SQL before aggregating

Example fix

// before
SELECT APPROX_COUNT_DISTINCT_DS_HLL(weird_col) FROM t; -- weird_col typed OTHER
// after
SELECT APPROX_COUNT_DISTINCT_DS_HLL(CAST(weird_col AS VARCHAR)) FROM t;

Defensive patterns

Strategy: validation

Validate before calling

ColumnType t = Calcites.getColumnTypeForRelDataType(relDataType); if (t == null) { throw new IllegalArgumentException("unsupported SQL type for sketch field: " + relDataType.getSqlTypeName()); }

Try / catch

try { agg = aggregator.toDruidAggregation(name, metricFactory, temporaryTable, input); } catch (IllegalStateException e) { throw new IllegalStateException("Check input column type for sketch aggregation: " + e.getMessage(), e); }

Prevention

Aggregate only over string-typed base64 sketch columns or numeric columns
Inspect segment schema (OTHER/NULL types) before writing sketch SQL
Cast unsupported expressions to VARCHAR/numeric in SQL first

When it happens

Trigger: Running a sketch aggregation whose first argument is a column of an unmappable SQL type (e.g. NULL, OTHER, complex/custom types) rather than a string-typed base64 sketch or a numeric column to count-distinguish.

Common situations: Sketch aggregation over a column typed as OTHER/complex in the segment metadata; passing a literal or expression with an unsupported RelDataType; schema drift where a column changed to an unsupported type.

Understand the failure class

Background: Type mismatch errors: IllegalArgumentException, TypeError and type guards across 150 open-source libraries — this error's family across 150 libraries.

Related errors

AI-assisted analysis of apache/druid@9b90983fd2 (2026-09-07). Data as JSON: /api/errors/c9bc5581ce11ece6. Report an issue: GitHub.