RecordBatch
This documentation reflects Arrow JS v4.0. Needs to be updated for the new Arrow API in v9.0 +.
Overview
A Record Batch in Apache Arrow is a collection of equal-length array instances.
Usage
A record batch can be created from this list of arrays using RecordBatch.from
:
const data = [
new Array([1, 2, 3, 4]),
new Array(['foo', 'bar', 'baz', None]),
new Array([True, None, False, True])
]
const recordBatch = RecordBatch.from(arrays);
Inheritance
RecordBatch
extends StructVector
extends BaseVector
Members
schema : Schema (readonly)
Returns the schema of the data in the record batch
numCols : Number (readonly)
Returns number of fields/columns in the schema (shorthand for this.schema.fields.length
).
Static Methods
RecordBatch.from(vectors: Array, names: String[] = []) : RecordBatch
Creates a RecordBatch
, see RecordBatch.new()
.
RecordBatch.new(vectors: Array, names: String[] = []) : RecordBatch
Creates new a record batch.
Schema is auto inferred, using names or index positions if names
are not supplied.
Methods
constructor(schema: Schema, numRows: Number, childData: (Data | Vector)[])
Create a new RecordBatch
instance with numRows
rows of child data.
numRows
-childData
-
constructor(schema: Schema, data: Data, children?: Vector[])
Create a new RecordBatch
instance with numRows
rows of child data.
constructor(...args: any[])
clone(data: Data, children?: Array) : RecordBatch
Returns a newly allocated copy of this RecordBatch
concat(...others: Vector[]) : Table
Concatenates a number of Vector
instances.
select(...columnNames: K[]) : RecordBatch
Return a new RecordBatch
with a subset of columns.