Implementation:Apache Flink FileSourceReader Constructor

Knowledge Sources	Apache Flink
Domains	Stream_Processing, File_IO
Last Updated	2026-02-09 00:00 GMT

Overview

Concrete tool for reading records from file splits using the single-threaded multiplexed reader pattern provided by the Apache Flink connector-files module.

Description

The FileSourceReader extends SingleThreadMultiplexSourceReaderBase to read file splits. Its constructor wires together FileSourceSplitReader (for file I/O via BulkFormat.Reader) and FileSourceRecordEmitter (for emitting records and tracking offsets). On startup, if no splits are assigned, it immediately requests one. When a split finishes, it requests the next. The reader is generic over split types via SplitT extends FileSourceSplit.

Usage

This is an internal class created by FileSource.createReader(). Users interact with it indirectly through the DataStream API.

Code Reference

Source Location

Repository: Apache Flink
File: flink-connectors/flink-connector-files/src/main/java/org/apache/flink/connector/file/src/impl/FileSourceReader.java
Lines: L35-72

Signature

@Internal
public final class FileSourceReader<T, SplitT extends FileSourceSplit>
        extends SingleThreadMultiplexSourceReaderBase<
                RecordAndPosition<T>, T, SplitT, FileSourceSplitState<SplitT>> {

    public FileSourceReader(
            SourceReaderContext readerContext,
            BulkFormat<T, SplitT> readerFormat,
            Configuration config) {
        super(
                () -> new FileSourceSplitReader<>(config, readerFormat),
                new FileSourceRecordEmitter<>(),
                config,
                readerContext);
    }

    @Override
    public void start() {
        if (getNumberOfCurrentlyAssignedSplits() == 0) {
            context.sendSplitRequest();
        }
    }

    @Override
    protected void onSplitFinished(Map<String, FileSourceSplitState<SplitT>> finishedSplitIds) {
        context.sendSplitRequest();
    }

    @Override
    protected FileSourceSplitState<SplitT> initializedState(SplitT split) {
        return new FileSourceSplitState<>(split);
    }

    @Override
    protected SplitT toSplitType(String splitId, FileSourceSplitState<SplitT> splitState) {
        return splitState.toFileSourceSplit();
    }
}

Import

import org.apache.flink.connector.file.src.impl.FileSourceReader;
// Internal class

I/O Contract

Inputs

Name	Type	Required	Description
readerContext	SourceReaderContext	Yes	Runtime context for split requests
readerFormat	BulkFormat<T, SplitT>	Yes	Format for reading records from splits
config	Configuration	Yes	Reader configuration

Outputs

Name	Type	Description
records	T	Individual records emitted downstream
split requests	SourceEvent	Requests for new splits sent to enumerator

Usage Examples

Reader Lifecycle

// FileSourceReader lifecycle:
// 1. start() -> sends split request if no splits assigned
// 2. Receives split from enumerator
// 3. FileSourceSplitReader.fetch() reads batches from the split
// 4. FileSourceRecordEmitter emits records and updates offset state
// 5. onSplitFinished() -> requests next split
// 6. Repeat until no more splits

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment