org.apache.directory.server.core.partition.impl.btree.jdbm

Class JdbmTable<K,V>

java.lang.Object

org.apache.directory.server.xdbm.AbstractTable<K,V>

org.apache.directory.server.core.partition.impl.btree.jdbm.JdbmTable<K,V>

All Implemented Interfaces:

Table<K,V>

Direct Known Subclasses:

JdbmMasterTable

public class JdbmTable<K,V>
extends AbstractTable<K,V>

A jdbm Btree wrapper that enables duplicate sorted keys using collections.

Field Summary

Fields

Modifier and Type	Field and Description
private jdbm.btree.BTree<K,V>	bt the wrappedCursor JDBM btree used in this Table
private java.util.Map<java.lang.Long,jdbm.btree.BTree<K,V>>	duplicateBtrees a cache of duplicate BTrees
private jdbm.helper.Serializer	keySerializer A Key serializer
private static org.slf4j.Logger	LOG A logger for this class
(package private) Marshaller<ArrayTree<V>>	marshaller A marshaller used to serialize/deserialize values stored in the Table
private int	numDupLimit the limit at which we start using btree redirection for duplicates
private jdbm.RecordManager	recMan the JDBM record manager for the file this table is managed in
private static java.lang.String	SZSUFFIX the key to store and retreive the count information
private jdbm.helper.Serializer	valueSerializer A value serializer

Fields inherited from class org.apache.directory.server.xdbm.AbstractTable

allowsDuplicates, commitNumber, count, keyComparator, name, schemaManager, valueComparator

Constructor Summary

Constructors

Constructor and Description
JdbmTable(org.apache.directory.api.ldap.model.schema.SchemaManager schemaManager, java.lang.String name, int numDupLimit, jdbm.RecordManager manager, java.util.Comparator<K> keyComparator, java.util.Comparator<V> valueComparator, jdbm.helper.Serializer keySerializer, jdbm.helper.Serializer valueSerializer) Creates a Jdbm BTree based tuple Table abstraction that enables duplicates.
JdbmTable(org.apache.directory.api.ldap.model.schema.SchemaManager schemaManager, java.lang.String name, jdbm.RecordManager manager, java.util.Comparator<K> keyComparator, jdbm.helper.Serializer keySerializer, jdbm.helper.Serializer valueSerializer) Creates a Jdbm BTree based tuple Table abstraction without duplicates enabled using a simple key comparator.

Method Summary

Modifier and Type	Method and Description
private boolean	btreeHas(jdbm.btree.BTree tree, V key, boolean isGreaterThan)
void	close() Closes the underlying Db of this Table.
private void	commit(jdbm.RecordManager recordManager) Commit the modification on disk
private ArrayTree<V>	convertToArrayTree(jdbm.btree.BTree bTree)
private jdbm.btree.BTree<V,K>	convertToBTree(ArrayTree<V> arrayTree)
long	count(K key) Gets the count of the number of records in this Table with a specific key: returns the number of duplicates for a key.
org.apache.directory.api.ldap.model.cursor.Cursor<org.apache.directory.api.ldap.model.cursor.Tuple<K,V>>	cursor() Creates a Cursor that traverses Tuples in a Table.
org.apache.directory.api.ldap.model.cursor.Cursor<org.apache.directory.api.ldap.model.cursor.Tuple<K,V>>	cursor(K key) Creates a Cursor that traverses Table Tuples for the same key.
V	get(K key) Gets the value of a record by key if the key exists.
(package private) jdbm.btree.BTree	getBTree() Returns the main BTree used by this table.
(package private) jdbm.btree.BTree	getBTree(BTreeRedirect redirect)
(package private) DupsContainer<V>	getDupsContainer(byte[] serialized)
jdbm.helper.Serializer	getKeySerializer()
Marshaller<ArrayTree<V>>	getMarshaller()
jdbm.helper.Serializer	getValueSerializer()
long	greaterThanCount(K key) Gets the number of records greater than or equal to a key value.
boolean	has(K key) Checks to see if this table has one or more tuples with a specific key: this is exactly the same as a get call with a check to see if the returned value is null or not.
boolean	has(K key, V value) Checks to see if this table has a key with a specific value.
boolean	hasGreaterOrEqual(K key) Checks to see if this table has a record with a key greater than or equal to the key argument.
boolean	hasGreaterOrEqual(K key, V val) Checks to see if this table has a Tuple with a key equal to the key argument, yet with a value greater than or equal to the value argument provided.
boolean	hasLessOrEqual(K key) Checks to see if this table has a record with a key less than or equal to the key argument.
boolean	hasLessOrEqual(K key, V val) Checks to see if this table has a Tuple with a key equal to the key argument, yet with a value less than or equal to the value argument provided.
boolean	isDupsEnabled() Checks to see if this Table has allows for duplicate keys (a.k.a.
(package private) boolean	isKeyUsingBTree(K key) Added to check that we actually switch from one data structure to the B+Tree structure on disk for duplicates that go beyond the threshold.
long	lessThanCount(K key) Gets the number of records less than or equal to a key value.
void	put(K key, V value) Puts a record into this Table.
void	remove(K key) Removes all records with a specified key from this Table.
void	remove(K key, V value) Removes a single key value pair with a specified key and value from this Table.
void	sync() Synchronizes the buffers with disk.
org.apache.directory.api.ldap.model.cursor.Cursor<V>	valueCursor(K key) Creates a Cursor that traverses Table values for the same key.

Methods inherited from class org.apache.directory.server.xdbm.AbstractTable

count, getKeyComparator, getName, getValueComparator, toString

Methods inherited from class java.lang.Object

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

Field Detail

LOG

private static final org.slf4j.Logger LOG

A logger for this class

SZSUFFIX

private static final java.lang.String SZSUFFIX

the key to store and retreive the count information

See Also:

Constant Field Values

recMan

private final jdbm.RecordManager recMan

the JDBM record manager for the file this table is managed in

bt

private jdbm.btree.BTree<K,V> bt

the wrappedCursor JDBM btree used in this Table

numDupLimit

private int numDupLimit

the limit at which we start using btree redirection for duplicates

duplicateBtrees

private final java.util.Map<java.lang.Long,jdbm.btree.BTree<K,V>> duplicateBtrees

a cache of duplicate BTrees

keySerializer

private final jdbm.helper.Serializer keySerializer

A Key serializer

valueSerializer

private final jdbm.helper.Serializer valueSerializer

A value serializer

marshaller

Marshaller<ArrayTree<V>> marshaller

A marshaller used to serialize/deserialize values stored in the Table

Constructor Detail

JdbmTable

public JdbmTable(org.apache.directory.api.ldap.model.schema.SchemaManager schemaManager,
                 java.lang.String name,
                 int numDupLimit,
                 jdbm.RecordManager manager,
                 java.util.Comparator<K> keyComparator,
                 java.util.Comparator<V> valueComparator,
                 jdbm.helper.Serializer keySerializer,
                 jdbm.helper.Serializer valueSerializer)
          throws java.io.IOException

Creates a Jdbm BTree based tuple Table abstraction that enables duplicates.

Parameters:

schemaManager - The server schemaManager

name - the name of the table

numDupLimit - the size limit of duplicates before switching to BTrees for values instead of AvlTrees

manager - the record manager to be used for this table

keyComparator - a key comparator

valueComparator - a value comparator

keySerializer - a serializer to use for the keys instead of using default Java serialization which could be very expensive

valueSerializer - a serializer to use for the values instead of using default Java serialization which could be very expensive

Throws:

java.io.IOException - if the table's file cannot be created

JdbmTable

public JdbmTable(org.apache.directory.api.ldap.model.schema.SchemaManager schemaManager,
                 java.lang.String name,
                 jdbm.RecordManager manager,
                 java.util.Comparator<K> keyComparator,
                 jdbm.helper.Serializer keySerializer,
                 jdbm.helper.Serializer valueSerializer)
          throws java.io.IOException

Creates a Jdbm BTree based tuple Table abstraction without duplicates enabled using a simple key comparator.

Parameters:

schemaManager - The server schemaManager

name - the name of the table

manager - the record manager to be used for this table

keyComparator - a tuple comparator

keySerializer - a serializer to use for the keys instead of using default Java serialization which could be very expensive

valueSerializer - a serializer to use for the values instead of using default Java serialization which could be very expensive

Throws:

java.io.IOException - if the table's file cannot be created

Method Detail

getKeySerializer

public jdbm.helper.Serializer getKeySerializer()

getValueSerializer

public jdbm.helper.Serializer getValueSerializer()

isDupsEnabled

public boolean isDupsEnabled()

Description copied from interface: Table

Checks to see if this Table has allows for duplicate keys (a.k.a. multiple values for the same key).

Returns:

true if duplicate keys are enabled, false otherwise

See Also:

Table.isDupsEnabled()

greaterThanCount

public long greaterThanCount(K key)
                      throws java.io.IOException

Description copied from interface: Table

Gets the number of records greater than or equal to a key value. The specific key argument provided need not exist for this call to return a non-zero value.

Parameters:

key - the key to use in comparisons

Returns:

the number of keys greater than or equal to the key

Throws:

java.io.IOException

See Also:

Table#greaterThanCount(Object)

lessThanCount

public long lessThanCount(K key)
                   throws java.io.IOException

Description copied from interface: Table

Gets the number of records less than or equal to a key value. The specific key argument provided need not exist for this call to return a non-zero value.

Parameters:

key - the key to use in comparisons

Returns:

the number of keys less than or equal to the key

Throws:

java.io.IOException

See Also:

Table#lessThanCount(Object)

count

public long count(K key)
           throws org.apache.directory.api.ldap.model.exception.LdapException

Description copied from interface: Table

Gets the count of the number of records in this Table with a specific key: returns the number of duplicates for a key.

Parameters:

key - the Object key to count.

Returns:

the number of duplicate records for a key.

Throws:

org.apache.directory.api.ldap.model.exception.LdapException

See Also:

Table.count(java.lang.Object)

get

public V get(K key)
      throws org.apache.directory.api.ldap.model.exception.LdapException

Description copied from interface: Table

Gets the value of a record by key if the key exists. If this Table allows duplicate keys then the first key will be returned. If this Table sorts keys then the key will be the smallest key in the Table as specificed by this Table's comparator or the default bytewise lexical comparator.

Parameters:

key - the key of the record

Returns:

the value of the record with the specified key if key exists or null if no such key exists.

Throws:

org.apache.directory.api.ldap.model.exception.LdapException

hasGreaterOrEqual

public boolean hasGreaterOrEqual(K key,
                                 V val)
                          throws org.apache.directory.api.ldap.model.exception.LdapException

Description copied from interface: Table

Checks to see if this table has a Tuple with a key equal to the key argument, yet with a value greater than or equal to the value argument provided. The key argument MUST exist for this call to return true and the underlying Db must allow for values of duplicate keys to be sorted. The entire basis to this method depends on the fact that tuples of the same key have values sorted according to a valid value comparator. If the table does not support duplicates then an UnsupportedOperationException is thrown.

Parameters:

key - the key

val - the value to compare values to

Returns:

true if a Tuple with a key equal to the key argument and a value greater than the value argument exists, false otherwise

Throws:

org.apache.directory.api.ldap.model.exception.LdapException

See Also:

Table#hasGreaterOrEqual(Object,Object)

hasLessOrEqual

public boolean hasLessOrEqual(K key,
                              V val)
                       throws org.apache.directory.api.ldap.model.exception.LdapException

Description copied from interface: Table

Checks to see if this table has a Tuple with a key equal to the key argument, yet with a value less than or equal to the value argument provided. The key argument MUST exist for this call to return true and the underlying Db must allow for values of duplicate keys to be sorted. The entire basis to this method depends on the fact that tuples of the same key have values sorted according to a valid value comparator. If the table does not support duplicates then an UnsupportedOperationException is thrown.

Parameters:

key - the key

val - the value to compare values to

Returns:

true if a Tuple with a key equal to the key argument and a value less than the value argument exists, false otherwise

Throws:

org.apache.directory.api.ldap.model.exception.LdapException

See Also:

Table#hasLessOrEqual(Object,Object)

hasGreaterOrEqual

public boolean hasGreaterOrEqual(K key)
                          throws java.io.IOException

Description copied from interface: Table

Checks to see if this table has a record with a key greater than or equal to the key argument. The key argument need not exist for this call to return true. The underlying database must sort keys based on a key comparator because this method depends on key ordering.

Parameters:

key - the key to compare keys to

Returns:

true if a Tuple with a key greater than or equal to the key argument exists, false otherwise

Throws:

java.io.IOException

See Also:

Table.hasGreaterOrEqual(Object)

hasLessOrEqual

public boolean hasLessOrEqual(K key)
                       throws java.io.IOException

Description copied from interface: Table

Checks to see if this table has a record with a key less than or equal to the key argument. The key argument need not exist for this call to return true. The underlying database must sort keys based on a key comparator because this method depends on key ordering.

Parameters:

key - the key to compare keys to

Returns:

true if a Tuple with a key less than or equal to the key argument exists, false otherwise

Throws:

java.io.IOException

See Also:

Table#hasLessOrEqual(Object)

has

public boolean has(K key,
                   V value)
            throws org.apache.directory.api.ldap.model.exception.LdapException

Description copied from interface: Table

Checks to see if this table has a key with a specific value.

Parameters:

key - the key to check for

value - the value to check for

Returns:

true if a record with the key and value exists, false otherwise

Throws:

org.apache.directory.api.ldap.model.exception.LdapException

See Also:

Table.has(java.lang.Object, java.lang.Object)

has

public boolean has(K key)
            throws java.io.IOException

Description copied from interface: Table

Checks to see if this table has one or more tuples with a specific key: this is exactly the same as a get call with a check to see if the returned value is null or not.

Parameters:

key - the Object of the key to check for

Returns:

true if the key exists, false otherwise

Throws:

java.io.IOException

See Also:

Table#has(java.lang.Object)

put

public void put(K key,
                V value)
         throws java.lang.Exception

Description copied from interface: Table

Puts a record into this Table. Null is not allowed for keys or values and should result in an IllegalArgumentException.

Parameters:

key - the key of the record

value - the value of the record.

Throws:

java.lang.Exception - if there is a failure to read or write to the underlying Db

java.lang.IllegalArgumentException - if a null key or value is used

See Also:

Table.put(java.lang.Object, java.lang.Object)

remove

public void remove(K key,
                   V value)
            throws java.io.IOException

Description copied from interface: Table

Removes a single key value pair with a specified key and value from this Table.

Parameters:

key - the key of the record to remove

value - the value of the record to remove

Throws:

java.io.IOException

See Also:

Table.remove(java.lang.Object, java.lang.Object)

remove

public void remove(K key)

Description copied from interface: Table

Removes all records with a specified key from this Table.

Parameters:

key - the key of the records to remove

See Also:

Table#remove(Object)

cursor

public org.apache.directory.api.ldap.model.cursor.Cursor<org.apache.directory.api.ldap.model.cursor.Tuple<K,V>> cursor()
                                                                                                                throws org.apache.directory.api.ldap.model.exception.LdapException

Description copied from interface: Table

Creates a Cursor that traverses Tuples in a Table.

Returns:

a Cursor over Tuples containing the key value pairs

Throws:

org.apache.directory.api.ldap.model.exception.LdapException

cursor

public org.apache.directory.api.ldap.model.cursor.Cursor<org.apache.directory.api.ldap.model.cursor.Tuple<K,V>> cursor(K key)
                                                                                                                throws java.lang.Exception

Description copied from interface: Table

Creates a Cursor that traverses Table Tuples for the same key. Only Tuples with the provided key will be returned if the key exists at all. If the key does not exist an empty Cursor is returned. The motivation behind this method is to minimize the need for callers to actively constrain Cursor operations based on the Tuples they return to a specific key. This Cursor is naturally limited to return only the tuples for the same key.

Parameters:

key - the duplicate key to return the Tuples of

Returns:

a Cursor over Tuples containing the same key

Throws:

java.lang.Exception - if there are failures accessing underlying stores

valueCursor

public org.apache.directory.api.ldap.model.cursor.Cursor<V> valueCursor(K key)
                                                                 throws java.lang.Exception

Description copied from interface: Table

Creates a Cursor that traverses Table values for the same key. Only Tuples with the provided key will have their values returned if the key exists at all. If the key does not exist an empty Cursor is returned. The motivation behind this method is to minimize the need for callers to actively constrain Cursor operations to a specific key while removing overheads in creating new Tuples or population one that is reused to return key value pairs. This Cursor is naturally limited to return only the values for the same key.

Parameters:

key - the duplicate key to return the values of

Returns:

a Cursor over values of a key

Throws:

java.lang.Exception - if there are failures accessing underlying stores

close

public void close()
           throws java.io.IOException

Description copied from interface: Table

Closes the underlying Db of this Table.

Throws:

java.io.IOException

See Also:

Table#close()

sync

public void sync()
          throws java.io.IOException

Synchronizes the buffers with disk.

Throws:

java.io.IOException - if errors are encountered on the flush

getMarshaller

public Marshaller<ArrayTree<V>> getMarshaller()

isKeyUsingBTree

boolean isKeyUsingBTree(K key)
                 throws java.lang.Exception

Added to check that we actually switch from one data structure to the B+Tree structure on disk for duplicates that go beyond the threshold.

Throws:

java.lang.Exception

getDupsContainer

DupsContainer<V> getDupsContainer(byte[] serialized)
                           throws org.apache.directory.api.ldap.model.exception.LdapException

Throws:

org.apache.directory.api.ldap.model.exception.LdapException

getBTree

jdbm.btree.BTree getBTree()

Returns the main BTree used by this table.

Returns:

the main JDBM BTree used by this table

getBTree

jdbm.btree.BTree getBTree(BTreeRedirect redirect)
                   throws java.io.IOException

Throws:

java.io.IOException

btreeHas

private boolean btreeHas(jdbm.btree.BTree tree,
                         V key,
                         boolean isGreaterThan)
                  throws java.io.IOException

Throws:

java.io.IOException

convertToArrayTree

private ArrayTree<V> convertToArrayTree(jdbm.btree.BTree bTree)
                                 throws java.io.IOException

Throws:

java.io.IOException

convertToBTree

private jdbm.btree.BTree<V,K> convertToBTree(ArrayTree<V> arrayTree)
                                      throws java.lang.Exception

Throws:

java.lang.Exception

commit

private void commit(jdbm.RecordManager recordManager)
             throws java.io.IOException

Commit the modification on disk

Parameters:

recordManager - The recordManager used for the commit

Throws:

java.io.IOException