Struct CollectArrowBridge 

pub struct CollectArrowBridge { /* private fields */ }

Auto-flushing bridge from row-level data collection to Arrow batches.

Rows are pushed via push_row. When the internal builder reaches batch_size, it is automatically flushed to the internal batch buffer. Call take_batches to retrieve all completed batches (including any partial data), or drain_completed to retrieve only full batches without flushing partial data.

Implementations

impl CollectArrowBridge

pub fn new(schema: SchemaRef, batch_size: usize) -> Result<Self, BridgeError>

Create a new bridge.

Arguments

• schema - Arrow schema for the batches.
• batch_size - number of rows per batch before auto-flush.

pub fn push_row(&mut self, values: &[ArrowValue]) -> Result<(), BridgeError>

Append a row. Auto-flushes to a new batch when batch_size is reached.

pub fn flush(&mut self) -> Result<(), BridgeError>

Flush any partial data in the builder to a batch.

pub fn take_batches(&mut self) -> Result<Vec<RecordBatch>, BridgeError>

Flush and return all accumulated batches (including partial data).

pub fn drain_completed(&mut self) -> Vec<RecordBatch>

Take only batches that have already been auto-flushed, without flushing partial data.

pub fn restore_completed(&mut self, batches: Vec<RecordBatch>)

Restore completed batches back into the bridge in their original order.

This is used when a downstream delivery stage fails after batches have already been drained from the bridge but before they were durably handed off to the next stage.

pub fn batches(&self) -> &[RecordBatch]

Borrow the accumulated (auto-flushed) batches.

pub fn pending_rows(&self) -> usize

Number of rows currently buffered but not yet flushed.

pub fn completed_batches(&self) -> usize

Number of completed batches waiting to be drained.

pub fn completed_rows(&self) -> usize

Number of rows contained in completed batches waiting to be drained.

pub fn batch_size(&self) -> usize

Configured auto-flush batch size.

pub fn metrics(&self) -> BridgeMetrics

Snapshot of current bridge metrics.

pub fn total_rows(&self) -> usize

Total row count (flushed + pending).