Package org.apache.flink.connector.file.src.enumerate

Class BlockSplittingRecursiveEnumerator

java.lang.Object

org.apache.flink.connector.file.src.enumerate.NonSplittingRecursiveEnumerator

org.apache.flink.connector.file.src.enumerate.BlockSplittingRecursiveEnumerator

All Implemented Interfaces:

FileEnumerator

Direct Known Subclasses:

BlockSplittingRecursiveAllDirEnumerator

@PublicEvolving
public class BlockSplittingRecursiveEnumerator
extends NonSplittingRecursiveEnumerator

This FileEnumerator enumerates all files under the given paths recursively, and creates a separate split for each file block.

Please note that file blocks are only exposed by some file systems, such as HDFS. File systems that do not expose block information will not create multiple file splits per file, but keep the files as one source split.

Files with suffixes corresponding to known compression formats (for example '.gzip', '.bz2', ...) will not be split. See StandardDeCompressors for a list of known formats and suffixes.

The default instantiation of this enumerator filters files with the common hidden file prefixes '.' and '_'. A custom file filter can be specified.

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.flink.connector.file.src.enumerate.FileEnumerator

FileEnumerator.Provider

Field Summary

Fields inherited from class org.apache.flink.connector.file.src.enumerate.NonSplittingRecursiveEnumerator

fileFilter

Constructor Summary

Constructors

Constructor	Description
BlockSplittingRecursiveEnumerator()	Creates a new enumerator that enumerates all files except hidden files.
BlockSplittingRecursiveEnumerator(Predicate<org.apache.flink.core.fs.Path> fileFilter, String[] nonSplittableFileSuffixes)	Creates a new enumerator that uses the given predicate as a filter for file paths, and avoids splitting files with the given extension (typically to avoid splitting compressed files).

Method Summary

Modifier and Type	Method	Description
protected void	convertToSourceSplits(org.apache.flink.core.fs.FileStatus file, org.apache.flink.core.fs.FileSystem fs, List<FileSourceSplit> target)
protected boolean	isFileSplittable(org.apache.flink.core.fs.Path filePath)

Methods inherited from class org.apache.flink.connector.file.src.enumerate.NonSplittingRecursiveEnumerator

addSplitsForPath, enumerateSplits, getNextId

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

BlockSplittingRecursiveEnumerator

public BlockSplittingRecursiveEnumerator()

Creates a new enumerator that enumerates all files except hidden files. Hidden files are considered files where the filename starts with '.' or with '_'.

The enumerator does not split files that have a suffix corresponding to a known compression format (for example '.gzip', '.bz2', '.xy', '.zip', ...). See StandardDeCompressors for details.

BlockSplittingRecursiveEnumerator

public BlockSplittingRecursiveEnumerator(Predicate<org.apache.flink.core.fs.Path> fileFilter,
 String[] nonSplittableFileSuffixes)

Creates a new enumerator that uses the given predicate as a filter for file paths, and avoids splitting files with the given extension (typically to avoid splitting compressed files).

Method Details

convertToSourceSplits

protected void convertToSourceSplits(org.apache.flink.core.fs.FileStatus file,
 org.apache.flink.core.fs.FileSystem fs,
 List<FileSourceSplit> target)
                              throws IOException

Overrides:

convertToSourceSplits in class NonSplittingRecursiveEnumerator

Throws:

IOException

isFileSplittable

protected boolean isFileSplittable(org.apache.flink.core.fs.Path filePath)