DataType

This enum specifies the data type of a collection field. Pass a DataType value when constructing a FieldSchema.

pub enum DataType {
    Unknown,
    Bool,
    Int8,
    Int16,
    Int32,
    Int64,
    Float,
    Double,
    VarChar,
    Json,
    Geometry,
    Timestamptz,
    Array,
    Struct,
    FloatVector,
    BinaryVector,
    Float16Vector,
    BFloat16Vector,
    SparseFloatVector,
    Int8Vector,
}

VARIANTS:

• Unknown

  Unspecified or unrecognized field type. This is the default.

• Bool

  Boolean value, true or false.

• Int8

  8-bit signed integer.

• Int16

  16-bit signed integer.

• Int32

  32-bit signed integer.

• Int64

  64-bit signed integer.

• Float

  32-bit IEEE 754 floating-point number.

• Double

  64-bit IEEE 754 floating-point number.

• VarChar

  Variable-length UTF-8 string. Constrain its size with max_length.

• Json

  JSON document stored as-is.

• Geometry

  Geometry value used by spatial queries (WKT-encoded).

• Timestamptz

  Timestamp with timezone.

• Array

  Array of scalar elements, all with the same element type.

• Struct

  Nested struct of named sub-fields.

• FloatVector

  Dense vector of 32-bit floats.

• BinaryVector

  Dense binary vector stored as packed bytes.

• Float16Vector

  Dense vector of IEEE 754 half-precision floats.

• BFloat16Vector

  Dense vector of bfloat16 values.

• SparseFloatVector

  Sparse vector stored as a map of dimension index to float value.

• Int8Vector

  Dense vector of 8-bit signed integers.

Example

let schema = CollectionSchema::new()
    .add_field(FieldSchema::new().name("id").data_type(DataType::Int64).primary_key(true))
    .add_field(FieldSchema::new().name("embedding").data_type(DataType::FloatVector).dimension(4));