All Classes and Interfaces
Class
Description
Class for a Binding-like structure except it works on NodeIds, not on Nodes
Bind that delays turning a NodeId into a Node until explicitly needed by get()
Interface to concrete storage - read and write Blocks, addressed by id.
Support for a disk file backed FileAccess
FileAccess interface backed by a byte array.
FileAccess for a file, using memory mapped I/O
File access layer that simulates a disk in-memory - for testing, not written for efficiency.
Caching block manager - this is an LRU cache
Block manager that maps from the FileAccess layer to a BlockMgr.
Recycle blocks - but only in-session.
Block manager that keeps temporary copies of updated blocks, then writes them
to a journal when commitPrepare happens.
Add synchronized to a BlockMgr.
B-Tree converted to B+Tree
B-Tree taken from:
Introduction to Algorithms, Second Edition
Chapter 18: B-Trees
by Thomas H.
Configuration for a B+Tree
BPlusTreePageMgr = BPlusTreeNode manager
Abstraction of a B+Tree node - either an branch (BTreeNode) or leaf (BTreeLeaf - records)
B+Tree wrapper over a block of records in a RecordBufferPage.
Bridge for making, getting and putting BPTreeRecords over a RecordBufferPageMgr
An allocator for retrieving ByteBuffers of a given size.
Delegates to
ByteBuffer.allocateDirect(int).ByteBuffer access to a temporary file, using memory mapped I/O.
Delegates to
ByteBuffer.allocate(int).Interface to storage : a simplified version of FileChannel.
Building datastructures on top of the base file
abstractions of indexes, block managers and object files.
Building indexes, blockMgr and object files
Overall framework for bulk loading
General descriptor of a reordering (mapping) of columns in tuples to columns in indexes,
for example, from triples to triple index order.
This class is the process of building a dataset.
A DatasetControl is an encapsulation of what to do on reads and writes.
A policy that checks, but does not enforce, single writer or multiple writer locking policy
Concurrency policy that is a "free for all".
A policy that provide read-only access
This is the class that creates a dataset over the storage.
A transactional
DatasetGraph that allows one active transaction per thread.A DatasetGraph that is a single transaction.
Dataset prefixes; a table of nodes with prefix-centric operations.
Abstract of prefix storage for graphs in an RDF dataset.
Lowlevel utilities for working with TDB
Extensible hashing
http://en.wikipedia.org/wiki/Extendible_hashing
File references.
Naming metadata management to a collection of related files
(same directory, same basename within directory, various extensions).
Non-transactional version of
GraphTDB.Projection from DatasetPrefixStorage to a single
PrefixMapping.General operations for TDB graphs
(free-standing graph, default graph and named graphs)
Transaction-capable version of
GraphTDB.Alternative GraphTxnTDB that has its own prefixes.
"switchable" prefix mapping backed by storage in DatasetPrefixesTDB
Hash values.
A HashBucket is a record buffer, with space store it's hash value and the bit length
Jena subsystem initialization
The Journal is slightly odd - it is append-only for write but random read.
Load into one NodeTupleTable (triples, quads, other)
Wrapper for a file system directory; can create filenames in that directory.
Represents a lock on a TDB location
Light weight (?) MRSW lock implementation that assumes
correct use of enterCriticalSection/leaveCriticalSection.
Abstraction and many convenience operations on metadata.
Names of things in TDB
Encode/decode for Nodes into bytes
Utilities for encoding/decoding nodes as strings.
Simple encoder/decoder for nodes that uses Turtle term string encoding.
Node table - conceptually a two way mapping of Node <-> NodeId
where Nodes can be stored and NodeIds allocated.
Cache wrapper around a NodeTable.
NodeTable wrapper to handle inline node ids.
A concrete NodeTable based on native storage (string file and an index)
Group a tuple table and node table together to provide a real NodeTupleTable
(Read-only?) projection of another NodeTupleTable.
An ObjectFile is an append-read file, that is you can append data
to the stream or read any block.
Variable length ByteBuffer file on disk with
buffering for delayed writes.
An ObjectFile is an append-read file, that is you can append data
to the stream or read any block.
This version works with two files - the main, base file and a per-transaction journal
file.
An ObjectFile is an append-read file, that is you can append data
to the stream or read any block.
TDB executor for algebra expressions.
A page with a byte buffer
Engine that wraps from blocks to typed pages.
Entry to the basic pattern solver for TDB1.
Single file, single ByteBuffer
Single, unsegmented file with ByteBuffer
From a file of records, build a (packed) index
Copy one index to another, probably with a different key order
Create the Node table and write the triples/quads temporary files
An IntBuffer with extra operations
Quad table - a collection of TupleIndexes for 4-tuples
together with a node table.
A record is pair of key and value.
B+Tree records nodes and hash buckets.
The on-disk form of a block of a single RecordBuffer
(i.e.
Manager for a block that is all records.
Record creator
Makes things : datasets from locations, indexes, etc etc.
Utilities used within the TDB BGP solver : local TDB store
RDF-star processing for matching an individual triple/quad pattern.
Execute TDB requests directly -- no reordering
Using OpExecutor is preferred.
Statistics collector, general purpose, uses Nodes
Statistics collector, aggregates based on NodeId
Statistics collector
A StoreConnection is the reference to the underlying storage.
System parameters for a TDB database instance.
Encode and decode
StoreParamsStore parameters that can be adjusted after a store has been created,
and given different values when the JVM attaches to a store area.
Wrap a
ObjectFile with a string encoder/decoder.Deprecated.
Backup a database.
Public factory for creating objects datasets backed by TDB storage
Public interface to the loader functionality.
Deprecated.
Use
TDB1BackupDeprecated.
Use
TDB1FactoryA collection of helpers to abstract away from internal details of TDB.
Deprecated.
Use
TDB1LoaderMake datasets in various ways.
Static information relating to TDB
A transaction.
Interface to stats for a single Transaction Manager
Interface that components of the transaction system implement.
TripleTable - a collection of TupleIndexes for 3-tuples
together with a node table.
A TupleTable is a set of TupleIndexes.
Buffered writer of rows of longs, in hex.
TDB1