apache/flink · error · InvalidTypesException

The given class is no subclass of {Writable.class.getName()}

Error message

The given class is no subclass of {Writable.class.getName()}

What it means

Thrown by the package-private factory WritableTypeInfo.getWritableTypeInfo(typeClass) when the supplied class is not an assignable subclass of Hadoop's Writable (or is exactly Writable.class). This is a type-extraction-time guard: only concrete Writable subclasses are admissible, and passing anything else raises InvalidTypesException. The message names the expected base class.

Source

Thrown at flink-connectors/flink-hadoop-compatibility/src/main/java/org/apache/flink/api/java/typeutils/WritableTypeInfo.java:155

        } else {
            return false;
        }
    }

    @Override
    public boolean canEqual(Object obj) {
        return obj instanceof WritableTypeInfo;
    }

    // --------------------------------------------------------------------------------------------

    @PublicEvolving
    static <T extends Writable> TypeInformation<T> getWritableTypeInfo(Class<T> typeClass) {
        if (Writable.class.isAssignableFrom(typeClass) && !typeClass.equals(Writable.class)) {
            return new WritableTypeInfo<T>(typeClass);
        } else {
            throw new InvalidTypesException(
                    "The given class is no subclass of " + Writable.class.getName());
        }
    }
}

View on GitHub (pinned to 2f3c205e92)

Solutions

1. Make the target class a concrete subclass of org.apache.hadoop.io.Writable (not the raw Writable type).
2. Ensure the class passed to WritableTypeInfo/getWritableTypeInfo actually implements Writable.
3. If you did not intend a Writable type, use the correct TypeInformation (POJO, generic, tuple) for that class.
4. Double-check imports so you pass your Writable subclass, not org.apache.hadoop.io.Writable itself.

Example fix

// before — passing the raw Writable interface
TypeInfo.of(org.apache.hadoop.io.Writable.class);
// after — pass a concrete Writable subclass
TypeInfo.of(MyWritable.class); // where MyWritable implements Writable

// Before requesting Writable type info, validate the class is a concrete Writable subclass
Class<?> c = candidateClass;
if (!org.apache.hadoop.io.Writable.class.isAssignableFrom(c) || c.equals(org.apache.hadoop.io.Writable.class)) {
    throw new IllegalArgumentException(c.getName()
        + " is not a concrete subclass of org.apache.hadoop.io.Writable");
}

org.apache.hadoop.io.Writable.class.isAssignableFrom(candidateClass) && !candidateClass.equals(org.apache.hadoop.io.Writable.class)

When it happens

Trigger: Produced when the type extractor resolves a type as a Writable and calls getWritableTypeInfo(typeClass) with a class that is not Writable-assignable or is the raw Writable class itself — e.g. annotating/declaring a type as Writable that is actually a non-Writable class, or requesting Writable type info for Writable.class directly.

Common situations: Declaring a POJO/non-Writable class where a Writable is expected; calling WritableTypeInfo on Writable.class (the base interface, not a subclass); a class that was Writable at build time but no longer is at runtime; misuse of the type-extraction API with a class that fails isAssignableFrom.

Related errors

• InputSplit must implement Writable interface.
• Cannot create Comparator for {typeClass.getCanonicalName()}.
• Hadoop input split must not be null
• Mapper may not be null.
• JobConf may not be null.

AI-assisted analysis of apache/flink@2f3c205e92 (2026-08-14). Data as JSON: /api/errors/e50f8e95d486cdaa. Report an issue: GitHub.