Package org.aksw.commons.io.hadoop.binseach.v2

Class BinarySearcherOverBlockSource

java.lang.Object

org.aksw.commons.io.hadoop.binseach.v2.BinarySearcherOverBlockSource

All Implemented Interfaces:

AutoCloseable, org.aksw.commons.io.binseach.BinarySearcher

public class BinarySearcherOverBlockSource
extends Object
implements org.aksw.commons.io.binseach.BinarySearcher

Binary search implementation that finds lines matching a prefix in a 'block' source such as a bzip2 compressed file.

Field Summary

Fields

Modifier and Type	Field	Description
protected BlockSource	blockSource
protected Supplier<org.aksw.commons.io.hadoop.binseach.v2.BinSearchResourceCache.CacheEntry>	cacheSupplier

Constructor Summary

Constructors

Constructor	Description
BinarySearcherOverBlockSource(BlockSource blockSource, Supplier<org.aksw.commons.io.hadoop.binseach.v2.BinSearchResourceCache.CacheEntry> cacheSupplier)

Method Summary

Modifier and Type	Method	Description
static long	adjustStart(BlockSource blockSource, long start, int depth, BinSearchLevelCache cache)	Adjust a position to the next block.
static Match	binarySearch(BlockSource blockSource, byte[] prefix, BinSearchLevelCache cache)
static Match	binarySearch(BlockSource blockSource, SearchMode searchMode, int depth, long start, long startAfter, long end, byte delimiter, byte[] prefix, BinSearchLevelCache cache)	When this method returns the input stream's position is unspecified.
void	close()
Stream<org.aksw.commons.io.input.ReadableChannelSupplier<byte[]>>	parallelSearch(byte[] prefix)
InputStream	search(byte[] prefix)

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.aksw.commons.io.binseach.BinarySearcher

search

Field Details

blockSource

protected BlockSource blockSource

cacheSupplier

protected Supplier<org.aksw.commons.io.hadoop.binseach.v2.BinSearchResourceCache.CacheEntry> cacheSupplier

Constructor Details

BinarySearcherOverBlockSource

public BinarySearcherOverBlockSource(BlockSource blockSource,
 Supplier<org.aksw.commons.io.hadoop.binseach.v2.BinSearchResourceCache.CacheEntry> cacheSupplier)

Method Details

close

public void close()
           throws Exception

Specified by:

close in interface AutoCloseable

Specified by:

close in interface org.aksw.commons.io.binseach.BinarySearcher

Throws:

Exception

search

public InputStream search(byte[] prefix)
                   throws IOException

Specified by:

search in interface org.aksw.commons.io.binseach.BinarySearcher

Throws:

IOException

parallelSearch

public Stream<org.aksw.commons.io.input.ReadableChannelSupplier<byte[]>> parallelSearch(byte[] prefix)
                                                                                 throws IOException

Specified by:

parallelSearch in interface org.aksw.commons.io.binseach.BinarySearcher

Throws:

IOException

binarySearch

public static Match binarySearch(BlockSource blockSource,
 byte[] prefix,
 BinSearchLevelCache cache)
                          throws IOException

Throws:

IOException

adjustStart

public static long adjustStart(BlockSource blockSource,
 long start,
 int depth,
 BinSearchLevelCache cache)
                        throws IOException

Adjust a position to the next block. It must hold that this function is idempotent: adjustStart(adjustStart(offset)) = adjustStart(offset)

Throws:

IOException

binarySearch

public static Match binarySearch(BlockSource blockSource,
 SearchMode searchMode,
 int depth,
 long start,
 long startAfter,
 long end,
 byte delimiter,
 byte[] prefix,
 BinSearchLevelCache cache)
                          throws IOException

When this method returns the input stream's position is unspecified.

Note on cache semantics:

1. Disposition is the mapping from the current offset to that of the next block (NOT the record).
2. The header map maps the block id to the first record

Parameters:

in - The seekable input stream on which to perform binary search for the given prefix.

searchMode - Whether we are searching the initial match, or the start or end of a run of matches.

start -

startAfter - The adjusted start for the offset (start + 1)

end -

delimiter -

prefix -

Returns:

Throws:

IOException