apache/iceberg · error

Unsupported type: + type

Error message

Unsupported type: + type

What it means

TimeTransform.fromSourceType maps the source column type to the per-type transform result: only DATE, TIMESTAMP, and TIMESTAMP_NANO sources are valid for time-based transforms (year/month/day/hour). Given any other input type — e.g. instant() (timestamptz depending on version), time, or a non-temporal type — it throws IllegalArgumentException because the transform is undefined for that source type.

Source

Thrown at api/src/main/java/org/apache/iceberg/transforms/TimeTransform.java:42

import org.apache.iceberg.expressions.UnboundPredicate;
import org.apache.iceberg.types.Type;
import org.apache.iceberg.util.SerializableFunction;

abstract class TimeTransform<S> implements Transform<S, Integer> {
  protected static <R> R fromSourceType(Type type, R dateResult, R microsResult, R nanosResult) {
    switch (type.typeId()) {
      case DATE:
        if (dateResult != null) {
          return dateResult;
        }
        break;
      case TIMESTAMP:
        return microsResult;
      case TIMESTAMP_NANO:
        return nanosResult;
    }

    throw new IllegalArgumentException("Unsupported type: " + type);
  }

  protected abstract ChronoUnit granularity();

  protected abstract Transform<S, Integer> toEnum(Type type);

  @Override
  public SerializableFunction<S, Integer> bind(Type type) {
    return toEnum(type).bind(type);
  }

  @Override
  public boolean preservesOrder() {
    return true;
  }

  @Override
  public boolean satisfiesOrderOf(Transform<?, ?> other) {

View on GitHub (pinned to 86d9c8fc54)

Solutions

1. Check the source column's type before applying a time transform: only date and timestamp (with/without zone, depending on version) are valid
2. Use the transform on a date or timestamp column in the partition spec, e.g. day(TimestampType.withoutZone()) not day(TimeType.get())
3. If the field type can vary, branch: use Identity or another transform for non-temporal types

Example fix

// before
Transform<Integer, Integer> t = Transforms.day(Types.TimeType.get()); // IllegalArgumentException
// after
Types.NestedField field = schema.findField("event_ts");
Transform<Integer, Integer> t = field.type() instanceof Types.TimestampType
    ? Transforms.day(field.type())
    : Transforms.identity();

boolean validTimeTransformSource(Type sourceType) {
  return sourceType.typeId() == Type.TypeID.DATE
      || sourceType.typeId() == Type.TypeID.TIMESTAMP
      || sourceType.typeId() == Type.TypeID.TIMESTAMP_NANO;
}
// verify the partition field's source column type before Transforms.day(t)/year(t)/etc.

if (!(type instanceof Types.DateType) && !(type instanceof Types.TimestampType)
    && !(type instanceof Types.TimestampNanoType)) {
  throw new IllegalArgumentException("Time transforms require a date or timestamp source: " + type);
}

try {
  Transform<Integer, Integer> t = Transforms.day(sourceType);
} catch (IllegalArgumentException e) {
  if (e.getMessage().startsWith("Unsupported type")) {
    throw new IllegalArgumentException("day() needs a date/timestamp column, got " + sourceType, e);
  }
  throw e;
}

When it happens

Trigger: Calling Transforms.year(t)/month(t)/day(t)/hour(t) (or toEnum) with a Type that is not DATE, TIMESTAMP, or TIMESTAMP_NANO — for example day(Types.TimeType.get()), day(Types.StringType.get()), or on some versions day(TimestampType.withZone()).

Common situations: Generic schema-translation code that blindly applies the same transform to whatever source type a partition field references; specs written against different Iceberg versions where timestamptz handling changed; typos where a time-type column is partitioned by day instead of a timestamp.

Understand the failure class

Background: "is not a compatible type" / "cannot merge" errors: when a value's type doesn't match what the library requires — this error's family across 65 libraries.

Related errors

• Cannot create expression literal from %s: %s
• Cannot bucket by type:
• hash(value) is not supported on the base Bucket class
• Unsupported time unit:
• Unsupported binary type: + value.getClass()

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/74d8cb14b50ed37f. Report an issue: GitHub.