Package org.apache.druid.segment.join.table

Class RowBasedIndexedTable<RowType>

java.lang.Object

org.apache.druid.segment.join.table.RowBasedIndexedTable<RowType>

All Implemented Interfaces:

Closeable, AutoCloseable, IndexedTable, ReferenceCountedObjectProvider<Closeable>

public class RowBasedIndexedTable<RowType>
extends Object
implements IndexedTable

An IndexedTable composed of a List-based table and Map-based indexes. The implementation is agnostic to the specific row type; it uses a RowAdapter to work with any sort of object. The class allows passing in a cache key. If the key is non-null, results of any join on this table can be cached. That cache becomes invalidated if this key changes. Creators of this class can pass in a non-null cache key if its possible to construct a small identifier - that must change when contents of this indexed table chances - May remain unchanged when contents of this indexed table How the cache key is constructed itself, depends on how the RowBasedIndexedTable is being built.

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.druid.segment.join.table.IndexedTable

IndexedTable.Index, IndexedTable.Reader

Constructor Summary

Constructors

Constructor	Description
RowBasedIndexedTable(List<RowType> table, RowAdapter<RowType> rowAdapter, RowSignature rowSignature, Set<String> keyColumns, String version)
RowBasedIndexedTable(List<RowType> table, RowAdapter<RowType> rowAdapter, RowSignature rowSignature, Set<String> keyColumns, String version, byte[] cacheKey)

Method Summary

Modifier and Type	Method	Description
Optional<Closeable>	acquireReference()	This method increments a reference count and provides a Closeable that decrements the reference count when closed.
void	close()
IndexedTable.Index	columnIndex(int column)	Returns the index for a particular column.
IndexedTable.Reader	columnReader(int column)	Returns a reader for a particular column.
byte[]	computeCacheKey()	Computes a byte[] key for the table that can be used for computing cache keys for join operations.
boolean	isCacheable()	Returns whether this indexed table can be cached for the join operations
Set<String>	keyColumns()	Returns the columns of this table that have indexes.
int	numRows()	Returns the number of rows in this table.
RowSignature	rowSignature()	Returns the signature of this table, which includes all key columns (as well as other columns that can be selected, but are not keys).
String	version()	Returns the version of this table, used to compare against when loading a new version of the table

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.druid.segment.join.table.IndexedTable

makeColumnSelectorFactory

Constructor Details

RowBasedIndexedTable

public RowBasedIndexedTable(List<RowType> table,
 RowAdapter<RowType> rowAdapter,
 RowSignature rowSignature,
 Set<String> keyColumns,
 String version)

RowBasedIndexedTable

public RowBasedIndexedTable(List<RowType> table,
 RowAdapter<RowType> rowAdapter,
 RowSignature rowSignature,
 Set<String> keyColumns,
 String version,
 @Nullable
 byte[] cacheKey)

Method Details

version

public String version()

Description copied from interface: IndexedTable

Returns the version of this table, used to compare against when loading a new version of the table

Specified by:

version in interface IndexedTable

keyColumns

public Set<String> keyColumns()

Description copied from interface: IndexedTable

Returns the columns of this table that have indexes.

Specified by:

keyColumns in interface IndexedTable

rowSignature

public RowSignature rowSignature()

Description copied from interface: IndexedTable

Returns the signature of this table, which includes all key columns (as well as other columns that can be selected, but are not keys).

Specified by:

rowSignature in interface IndexedTable

columnIndex

public IndexedTable.Index columnIndex(int column)

Description copied from interface: IndexedTable

Returns the index for a particular column. The provided column number must be that column's position in IndexedTable.rowSignature().

Specified by:

columnIndex in interface IndexedTable

columnReader

public IndexedTable.Reader columnReader(int column)

Description copied from interface: IndexedTable

Returns a reader for a particular column. The provided column number must be that column's position in IndexedTable.rowSignature(). Don't forget to close your IndexedTable.Reader when finished reading, to clean up any resources.

Specified by:

columnReader in interface IndexedTable

computeCacheKey

public byte[] computeCacheKey()

Description copied from interface: IndexedTable

Computes a byte[] key for the table that can be used for computing cache keys for join operations. see JoinableFactory.computeJoinCacheKey(org.apache.druid.query.DataSource, org.apache.druid.segment.join.JoinConditionAnalysis)

Specified by:

computeCacheKey in interface IndexedTable

Returns:

the byte array for cache key

isCacheable

public boolean isCacheable()

Description copied from interface: IndexedTable

Returns whether this indexed table can be cached for the join operations

Specified by:

isCacheable in interface IndexedTable

numRows

public int numRows()

Description copied from interface: IndexedTable

Returns the number of rows in this table. It must not change over time, since it is used for things like algorithm selection and reporting of cardinality metadata.

Specified by:

numRows in interface IndexedTable

acquireReference

public Optional<Closeable> acquireReference()

Description copied from interface: ReferenceCountedObjectProvider

This method increments a reference count and provides a Closeable that decrements the reference count when closed. The object returned by this method may or may not be thread-safe, it is best to check with the implementation.

IMPORTANT NOTE: to fulfill the contract of this method, implementors must return a closeable to indicate that the reference can be acquired, even if there is nothing to close. Implementors should ideally avoid allowing this method to throw exceptions, but callers are responsible for closing the returned object should any operations on it throw exceptions.

For callers: if this method returns non-empty, IT MUST BE CLOSED, else reference counts can potentially leak.

Specified by:

acquireReference in interface ReferenceCountedObjectProvider<RowType>

close

public void close()

Specified by:

close in interface AutoCloseable

Specified by:

close in interface Closeable