pub fn decode_records_sequential<T: DeserializeOwned, B: AsRef<[u8]>>(
schema: &Schema,
binary_records: &[B],
) -> Result<Vec<T>, KineticaError>Decodes binary Avro records sequentially using serde_avro_fast.
Use this for small result sets (<500 records) where parallel overhead would outweigh benefits. For automatic threshold handling, use decode_records.
Performance
Achieves 2-4M records/second, 3-5x faster than apache_avro.