Struct Encoder 

pub struct Encoder {
    schema: SchemaRef,
    encoder: RecordEncoder,
    row_capacity: Option<usize>,
    buffer: BytesMut,
    offsets: Vec<usize>,
}

A row-by-row encoder for Avro stream/message formats (SOE / registry wire formats / raw binary).

Unlike Writer, which emits a single continuous byte stream to a std::io::Write sink, Encoder tracks row boundaries during encoding and returns an EncodedRows containing:

• one backing buffer (Bytes)
• row boundary offsets

This enables zero-copy per-row payloads (for instance, one Kafka message per Arrow row) without re-encoding or decoding the byte stream to recover record boundaries.

Example

use std::sync::Arc;
use arrow_array::{ArrayRef, Int32Array, RecordBatch};
use arrow_schema::{DataType, Field, Schema};
use arrow_avro::writer::{WriterBuilder, format::AvroSoeFormat};
use arrow_avro::schema::FingerprintStrategy;

let schema = Schema::new(vec![Field::new("value", DataType::Int32, false)]);
let batch = RecordBatch::try_new(
    Arc::new(schema.clone()),
    vec![Arc::new(Int32Array::from(vec![1, 2, 3])) as ArrayRef],
)?;

// Configure the encoder (here: Confluent Wire Format with schema ID 100)
let mut encoder = WriterBuilder::new(schema)
    .with_fingerprint_strategy(FingerprintStrategy::Id(100))
    .build_encoder::<AvroSoeFormat>()?;

// Encode the batch
encoder.encode(&batch)?;

// Get the encoded rows
let rows = encoder.flush();

// Convert to owned Vec<u8> payloads (e.g., for a Kafka producer)
let payloads: Vec<Vec<u8>> = rows.iter().map(|row| row.to_vec()).collect();

assert_eq!(payloads.len(), 3);
assert_eq!(payloads[0][0], 0x00); // Magic byte

Fields

schema: SchemaRef

encoder: RecordEncoder

row_capacity: Option<usize>

buffer: BytesMut

offsets: Vec<usize>

Implementations

impl Encoder

pub fn encode(&mut self, batch: &RecordBatch) -> Result<(), AvroError>

Serialize one [RecordBatch] into the internal buffer.

pub fn encode_batches( &mut self, batches: &[RecordBatch], ) -> Result<(), AvroError>

A convenience method to write a slice of [RecordBatch] values.

pub fn flush(&mut self) -> EncodedRows

Drain and return all currently buffered encoded rows.

The returned EncodedRows provides per-row payloads as Bytes slices.

pub fn schema(&self) -> SchemaRef

Returns the Arrow schema used by this encoder.

The returned schema includes metadata with the Avro schema JSON under the avro.schema key.

pub fn buffered_len(&self) -> usize

Returns the number of encoded rows currently buffered.

Trait Implementations

impl Debug for Encoder

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.