apache/cassandra · error · InvalidRequestException

vectors may only have positive dimensions; given

Error message

vectors may only have positive dimensions; given %d

What it means

VectorType requires a strictly positive dimension when a vector type is constructed. The constructor validates the dimension parameter and throws InvalidRequestException if it is <= 0, because a vector with zero or negative elements is meaningless in CQL.

Solutions

1. Pass a positive dimension (>= 1) when constructing the vector type
2. Fix the CQL schema to declare vector<elementType, N> with N > 0
3. If dimension comes from a variable/config, validate it is > 0 before calling VectorType.getInstance

Example fix

// before
AbstractType<?> t = VectorType.getInstance(Int32Type.instance, 0);
// after
AbstractType<?> t = VectorType.getInstance(Int32Type.instance, 3);

if (dimension <= 0) throw new IllegalArgumentException("dimension must be > 0, got " + dimension);
AbstractType<?> t = VectorType.getInstance(elementType, dimension);

try { VectorType.getInstance(et, dim); } catch (InvalidRequestException e) { /* log and fail schema setup */ }

When it happens

Trigger: Calling VectorType.getInstance(elementType, dimension) with dimension = 0 or a negative number, or declaring a CQL type like vector<int, 0> which resolves to a VectorType construction with an invalid dimension.

Common situations: Typo or computed dimension passed as 0 (e.g. a config value or variable that defaulted to 0), schema generation code building vector types programmatically, or user-supplied DDL with vector<..., 0>.

Understand the failure class

Background: "Must be a positive integer", "Invalid value", "Unsupported": the invalid-argument-value error family, when a library rejects the value you pass — this error's family across 35 libraries.

Related errors

• ACCESS TO DATACENTERS operations not supported by…
• ALREADY_EXISTS
• Already exists
• Attempted to add float vector of dimension
• Because of this name

AI-assisted analysis of apache/cassandra@88fd0f6a0e (2026-09-10). Data as JSON: /api/errors/a7a4eb475457ec77. Report an issue: GitHub.

Appendix: source

Thrown at src/java/org/apache/cassandra/db/marshal/VectorType.java:91

        @Override
        public int hashCode()
        {
            return Objects.hash(type, dimension);
        }
    }
    @SuppressWarnings("rawtypes")
    private static final ConcurrentHashMap<Key, VectorType> instances = new ConcurrentHashMap<>();

    public final AbstractType<T> elementType;
    public final int dimension;
    private final TypeSerializer<T> elementSerializer;
    private final VectorSerializer serializer;

    private VectorType(AbstractType<T> elementType, int dimension)
    {
        super(ComparisonType.CUSTOM, valueLengthIfFixed(elementType, dimension));
        if (dimension <= 0)
            throw new InvalidRequestException(String.format("vectors may only have positive dimensions; given %d", dimension));
        this.elementType = elementType;
        this.dimension = dimension;
        this.elementSerializer = elementType.getSerializer();
        this.serializer = elementType.isValueLengthFixed() ?
                          new FixedLengthSerializer() :
                          new VariableLengthSerializer();
    }

    @SuppressWarnings("unchecked")
    public static <T> VectorType<T> getInstance(AbstractType<T> elements, int dimension)
    {
        Key key = new Key(elements, dimension);
        return instances.computeIfAbsent(key, Key::create);
    }

    public static VectorType<?> getInstance(TypeParser parser)
    {
        TypeParser.Vector v = parser.getVectorParameters();

View on GitHub (pinned to 88fd0f6a0e)