This class provides read/write access to a database. More...

#include <database.h>

Inheritance diagram for Xapian::WritableDatabase:

Public Member Functions
virtual	~WritableDatabase ()
	Destroy this handle on the database.

	WritableDatabase ()
	Create a WritableDatabase with no subdatabases.

	WritableDatabase (const std::string &path, int flags=0, int block_size=0)
	Open a database for update, automatically determining the database backend to use.

	WritableDatabase (const WritableDatabase &other)
	Copying is allowed.

void	operator= (const WritableDatabase &other)
	Assignment is allowed.

void	add_database (const WritableDatabase &other)
	Add shards from another WritableDatabase.

void	commit ()
	Commit any pending modifications made to the database.

void	flush ()
	Pre-1.1.0 name for commit().

void	begin_transaction (bool flushed=true)
	Begin a transaction.

void	commit_transaction ()
	Complete the transaction currently in progress.

void	cancel_transaction ()
	Abort the transaction currently in progress, discarding the pending modifications made to the database.

Xapian::docid	add_document (const Xapian::Document &document)
	Add a new document to the database.

void	delete_document (Xapian::docid did)
	Delete a document from the database.

void	delete_document (const std::string &unique_term)
	Delete any documents indexed by a term from the database.

void	replace_document (Xapian::docid did, const Xapian::Document &document)
	Replace a given document in the database.

Xapian::docid	replace_document (const std::string &unique_term, const Xapian::Document &document)
	Replace any documents matching a term.

void	add_spelling (const std::string &word, Xapian::termcount freqinc=1) const
	Add a word to the spelling dictionary.

void	remove_spelling (const std::string &word, Xapian::termcount freqdec=1) const
	Remove a word from the spelling dictionary.

void	add_synonym (const std::string &term, const std::string &synonym) const
	Add a synonym for a term.

void	remove_synonym (const std::string &term, const std::string &synonym) const
	Remove a synonym for a term.

void	clear_synonyms (const std::string &term) const
	Remove all synonyms for a term.

void	set_metadata (const std::string &key, const std::string &metadata)
	Set the user-specified metadata associated with a given key.

std::string	get_description () const
	Return a string describing this object.

Public Member Functions inherited from Xapian::Database
void	add_database (const Database &database)
	Add an existing database (or group of databases) to those accessed by this object.

size_t	size () const
	Return number of shards in this Database object.

	Database ()
	Create a Database with no databases in.

	Database (const std::string &path, int flags=0)
	Open a Database, automatically determining the database backend to use.

	Database (int fd, int flags=0)
	Open a single-file Database.

virtual	~Database ()
	Destroy this handle on the database.

	Database (const Database &other)
	Copying is allowed.

void	operator= (const Database &other)
	Assignment is allowed.

bool	reopen ()
	Re-open the database.

virtual void	close ()
	Close the database.

PostingIterator	postlist_begin (const std::string &tname) const
	An iterator pointing to the start of the postlist for a given term.

PostingIterator	postlist_end (const std::string &) const
	Corresponding end iterator to postlist_begin().

TermIterator	termlist_begin (Xapian::docid did) const
	An iterator pointing to the start of the termlist for a given document.

TermIterator	termlist_end (Xapian::docid) const
	Corresponding end iterator to termlist_begin().

bool	has_positions () const
	Does this database have any positional information?

PositionIterator	positionlist_begin (Xapian::docid did, const std::string &tname) const
	An iterator pointing to the start of the position list for a given term in a given document.

PositionIterator	positionlist_end (Xapian::docid, const std::string &) const
	Corresponding end iterator to positionlist_begin().

TermIterator	allterms_begin (const std::string &prefix=std::string()) const
	An iterator which runs across all terms with a given prefix.

TermIterator	allterms_end (const std::string &=std::string()) const
	Corresponding end iterator to allterms_begin(prefix).

Xapian::doccount	get_doccount () const
	Get the number of documents in the database.

Xapian::docid	get_lastdocid () const
	Get the highest document id which has been used in the database.

Xapian::doclength	get_avlength () const
	Get the average length of the documents in the database.

double	get_average_length () const
	New name for get_avlength().

Xapian::totallength	get_total_length () const
	Get the total length of all the documents in the database.

Xapian::doccount	get_termfreq (const std::string &tname) const
	Get the number of documents in the database indexed by a given term.

bool	term_exists (const std::string &tname) const
	Check if a given term exists in the database.

Xapian::termcount	get_collection_freq (const std::string &tname) const
	Return the total number of occurrences of the given term.

Xapian::doccount	get_value_freq (Xapian::valueno slot) const
	Return the frequency of a given value slot.

std::string	get_value_lower_bound (Xapian::valueno slot) const
	Get a lower bound on the values stored in the given value slot.

std::string	get_value_upper_bound (Xapian::valueno slot) const
	Get an upper bound on the values stored in the given value slot.

Xapian::termcount	get_doclength_lower_bound () const
	Get a lower bound on the length of a document in this DB.

Xapian::termcount	get_doclength_upper_bound () const
	Get an upper bound on the length of a document in this DB.

Xapian::termcount	get_wdf_upper_bound (const std::string &term) const
	Get an upper bound on the wdf of term term.

ValueIterator	valuestream_begin (Xapian::valueno slot) const
	Return an iterator over the value in slot slot for each document.

ValueIterator	valuestream_end (Xapian::valueno) const
	Return end iterator corresponding to valuestream_begin().

Xapian::termcount	get_doclength (Xapian::docid did) const
	Get the length of a document.

Xapian::termcount	get_unique_terms (Xapian::docid did) const
	Get the number of unique terms in document.

void	keep_alive ()
	Send a "keep-alive" to remote databases to stop them timing out.

Xapian::Document	get_document (Xapian::docid did) const
	Get a document from the database, given its document id.

Xapian::Document	get_document (Xapian::docid did, unsigned flags) const
	Get a document from the database, given its document id.

std::string	get_spelling_suggestion (const std::string &word, unsigned max_edit_distance=2) const
	Suggest a spelling correction.

Xapian::TermIterator	spellings_begin () const
	An iterator which returns all the spelling correction targets.

Xapian::TermIterator	spellings_end () const
	Corresponding end iterator to spellings_begin().

Xapian::TermIterator	synonyms_begin (const std::string &term) const
	An iterator which returns all the synonyms for a given term.

Xapian::TermIterator	synonyms_end (const std::string &) const
	Corresponding end iterator to synonyms_begin(term).

Xapian::TermIterator	synonym_keys_begin (const std::string &prefix=std::string()) const
	An iterator which returns all terms which have synonyms.

Xapian::TermIterator	synonym_keys_end (const std::string &=std::string()) const
	Corresponding end iterator to synonym_keys_begin(prefix).

std::string	get_metadata (const std::string &key) const
	Get the user-specified metadata associated with a given key.

Xapian::TermIterator	metadata_keys_begin (const std::string &prefix=std::string()) const
	An iterator which returns all user-specified metadata keys.

Xapian::TermIterator	metadata_keys_end (const std::string &=std::string()) const
	Corresponding end iterator to metadata_keys_begin().

std::string	get_uuid () const
	Get a UUID for the database.

bool	locked () const
	Test if this database is currently locked for writing.

Xapian::rev	get_revision () const
	Get the revision of the database.

void	compact (const std::string &output, unsigned flags=0, int block_size=0)
	Produce a compact version of this database.

void	compact (int fd, unsigned flags=0, int block_size=0)
	Produce a compact version of this database.

void	compact (const std::string &output, unsigned flags, int block_size, Xapian::Compactor &compactor)
	Produce a compact version of this database.

void	compact (int fd, unsigned flags, int block_size, Xapian::Compactor &compactor)
	Produce a compact version of this database.

Additional Inherited Members
Static Public Member Functions inherited from Xapian::Database
static size_t	check (const std::string &path, int opts=0, std::ostream *out=NULL)
	Check the integrity of a database or database table.

static size_t	check (int fd, int opts=0, std::ostream *out=NULL)
	Check the integrity of a single file database.

Detailed Description

This class provides read/write access to a database.

Constructor & Destructor Documentation

◆ ~WritableDatabase()

virtual Xapian::WritableDatabase::~WritableDatabase ( )

virtual

Destroy this handle on the database.

If no other handles to this database remain, the database will be closed.

If a transaction is active cancel_transaction() will be implicitly called; if no transaction is active commit() will be implicitly called, but any exception will be swallowed (because throwing exceptions in C++ destructors is problematic). If you aren't using transactions and want to know about any failure to commit changes, call commit() explicitly before the destructor gets called.

◆ WritableDatabase() [1/3]

Xapian::WritableDatabase::WritableDatabase ( )

Create a WritableDatabase with no subdatabases.

The created object isn't very useful in this state - it's intended as a placeholder value.

◆ WritableDatabase() [2/3]

Xapian::WritableDatabase::WritableDatabase	(	const std::string &	path,
		int	flags = `0`,
		int	block_size = `0`
	)

explicit

Open a database for update, automatically determining the database backend to use.

If the database is to be created, Xapian will try to create the directory indicated by path if it doesn't already exist (but only the leaf directory, not recursively).

Parameters

path	directory that the database is stored in.
flags	one of: Xapian::DB_CREATE_OR_OPEN open for read/write; create if no db exists (the default if flags isn't specified) Xapian::DB_CREATE create new database; fail if db exists Xapian::DB_CREATE_OR_OVERWRITE overwrite existing db; create if none exists Xapian::DB_OPEN open for read/write; fail if no db exists

Additionally, the following flags can be combined with action using bitwise-or (| in C++):

Xapian::DB_NO_SYNC don't call fsync() or similar
Xapian::DB_FULL_SYNC try harder to ensure data is safe
Xapian::DB_DANGEROUS don't be crash-safe, no concurrent readers
Xapian::DB_NO_TERMLIST don't use a termlist table
Xapian::DB_RETRY_LOCK to wait to get a write lock

Parameters

block_size If a new database is created, this specifies the block size (in bytes) for backends which have such a concept. For chert and glass, the block size must be a power of 2 between 2048 and 65536 (inclusive), and the default (also used if an invalid value is passed) is 8192 bytes.

Exceptions

Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.
Xapian::DatabaseLockError	will be thrown if a lock couldn't be acquired on the database.

◆ WritableDatabase() [3/3]

Xapian::WritableDatabase::WritableDatabase ( const WritableDatabase & other )

Copying is allowed.

The internals are reference counted, so copying is cheap.

Parameters

other The object to copy.

Member Function Documentation

◆ add_database()

void Xapian::WritableDatabase::add_database ( const WritableDatabase & other )

inline

Add shards from another WritableDatabase.

Any shards in other are added to the list of shards in this object. The shards are reference counted and also remain in other.

Parameters

other Another WritableDatabase object to add shards from

◆ add_document()

Xapian::docid Xapian::WritableDatabase::add_document ( const Xapian::Document & document )

Add a new document to the database.

This method adds the specified document to the database, returning a newly allocated document ID. Automatically allocated document IDs come from a per-database monotonically increasing counter, so IDs from deleted documents won't be reused.

If you want to specify the document ID to be used, you should call replace_document() instead.

Note that changes to the database won't be immediately committed to disk; see commit() for more details.

As with all database modification operations, the effect is atomic: the document will either be fully added, or the document fails to be added and an exception is thrown (possibly at a later time when commit() is called or the database is closed).

Parameters

document The new document to be added.

Returns: The document ID of the newly added document.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while writing to the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.

◆ add_spelling()

void Xapian::WritableDatabase::add_spelling	(	const std::string &	word,
		Xapian::termcount	freqinc = `1`
	)		const

Add a word to the spelling dictionary.

If the word is already present, its frequency is increased.

Parameters

word	The word to add.
freqinc	How much to increase its frequency by (default 1).

◆ add_synonym()

void Xapian::WritableDatabase::add_synonym	(	const std::string &	term,
		const std::string &	synonym
	)		const

Add a synonym for a term.

Parameters

term	The term to add a synonym for.
synonym	The synonym to add. If this is already a synonym for term, then no action is taken.

◆ begin_transaction()

void Xapian::WritableDatabase::begin_transaction ( bool flushed = true )

Begin a transaction.

In Xapian a transaction is a group of modifications to the database which are linked such that either all will be applied simultaneously or none will be applied at all. Even in the case of a power failure, this characteristic should be preserved (as long as the filesystem isn't corrupted, etc).

However, note that if called on a sharded database, atomicity isn't guaranteed between shards. Within each shard, the transaction will still act atomically.

A transaction is started with begin_transaction() and can either be committed by calling commit_transaction() or aborted by calling cancel_transaction().

By default, a transaction implicitly calls commit() before and after so that the modifications stand and fall without affecting modifications before or after.

The downside of these implicit calls to commit() is that small transactions can harm indexing performance in the same way that explicitly calling commit() frequently can.

If you're applying atomic groups of changes and only wish to ensure that each group is either applied or not applied, then you can prevent the automatic commit() before and after the transaction by starting the transaction with begin_transaction(false). However, if cancel_transaction is called (or if commit_transaction isn't called before the WritableDatabase object is destroyed) then any changes which were pending before the transaction began will also be discarded.

Transactions aren't currently supported by the InMemory backend.

Parameters

flushed Is this a flushed transaction? By default transactions are "flushed", which means that committing a transaction will ensure those changes are permanently written to the database. By contrast, unflushed transactions only ensure that changes within the transaction are either all applied or all aren't.

Exceptions

Xapian::UnimplementedError	will be thrown if transactions are not available for this database type.
Xapian::InvalidOperationError	will be thrown if this is called at an invalid time, such as when a transaction is already in progress.

◆ cancel_transaction()

void Xapian::WritableDatabase::cancel_transaction ( )

Abort the transaction currently in progress, discarding the pending modifications made to the database.

If an error occurs in this method, an exception will be thrown, but the transaction will be cancelled anyway.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while modifying the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.
Xapian::InvalidOperationError	will be thrown if a transaction is not currently in progress.
Xapian::UnimplementedError	will be thrown if transactions are not available for this database type.

◆ clear_synonyms()

void Xapian::WritableDatabase::clear_synonyms ( const std::string & term ) const

Remove all synonyms for a term.

Parameters

term	The term to remove all synonyms for. If the term has no synonyms, no action is taken.

◆ commit()

void Xapian::WritableDatabase::commit ( )

Commit any pending modifications made to the database.

For efficiency reasons, when performing multiple updates to a database it is best (indeed, almost essential) to make as many modifications as memory will permit in a single pass through the database. To ensure this, Xapian batches up modifications.

This method may be called at any time to commit any pending modifications to the database.

If any of the modifications fail, an exception will be thrown and the database will be left in a state in which each separate addition, replacement or deletion operation has either been fully performed or not performed at all: it is then up to the application to work out which operations need to be repeated.

However, note that if called on a sharded database, atomicity isn't guaranteed between shards - it's possible for the changes to one shard to be committed but changes to another shard to fail.

It's not valid to call commit() within a transaction.

Beware of calling commit() too frequently: this will make indexing take much longer.

Note that commit() need not be called explicitly: it will be called automatically when the database is closed, or when a sufficient number of modifications have been made. By default, this is every 10000 documents added, deleted, or modified. This value is rather conservative, and if you have a machine with plenty of memory, you can improve indexing throughput dramatically by setting XAPIAN_FLUSH_THRESHOLD in the environment to a larger value.

This method was new in Xapian 1.1.0 - in earlier versions it was called flush().

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while modifying the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.

◆ commit_transaction()

void Xapian::WritableDatabase::commit_transaction ( )

Complete the transaction currently in progress.

If this method completes successfully and this is a flushed transaction, all the database modifications made during the transaction will have been committed to the database.

If an error occurs, an exception will be thrown, and none of the modifications made to the database during the transaction will have been applied to the database.

In all cases the transaction will no longer be in progress.

Note that if called on a sharded database, atomicity isn't guaranteed between shards. Within each shard, the transaction will still act atomically.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while modifying the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.
Xapian::InvalidOperationError	will be thrown if a transaction is not currently in progress.
Xapian::UnimplementedError	will be thrown if transactions are not available for this database type.

◆ delete_document() [1/2]

void Xapian::WritableDatabase::delete_document ( const std::string & unique_term )

Delete any documents indexed by a term from the database.

This method removes any documents indexed by the specified term from the database.

A major use is for convenience when UIDs from another system are mapped to terms in Xapian, although this method has other uses (for example, you could add a "deletion date" term to documents at index time and use this method to delete all documents due for deletion on a particular date).

Parameters

unique_term The term to remove references to.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while writing to the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.

◆ delete_document() [2/2]

void Xapian::WritableDatabase::delete_document ( Xapian::docid did )

Delete a document from the database.

This method removes the document with the specified document ID from the database.

Note that changes to the database won't be immediately committed to disk; see commit() for more details.

As with all database modification operations, the effect is atomic: the document will either be fully removed, or the document fails to be removed and an exception is thrown (possibly at a later time when commit() is called or the database is closed).

Parameters

did	The document ID of the document to be removed.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while writing to the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.

◆ flush()

void Xapian::WritableDatabase::flush ( )

inline

Pre-1.1.0 name for commit().

Use commit() instead.

◆ get_description()

std::string Xapian::WritableDatabase::get_description ( ) const

virtual

Return a string describing this object.

Reimplemented from Xapian::Database.

◆ operator=()

void Xapian::WritableDatabase::operator= ( const WritableDatabase & other )

Assignment is allowed.

The internals are reference counted, so assignment is cheap.

Note that only an WritableDatabase may be assigned to an WritableDatabase: an attempt to assign a Database is caught at compile-time.

Parameters

other The object to copy.

◆ remove_spelling()

void Xapian::WritableDatabase::remove_spelling	(	const std::string &	word,
		Xapian::termcount	freqdec = `1`
	)		const

Remove a word from the spelling dictionary.

The word's frequency is decreased, and if would become zero or less then the word is removed completely.

Parameters

word	The word to remove.
freqdec	How much to decrease its frequency by (default 1).

◆ remove_synonym()

void Xapian::WritableDatabase::remove_synonym	(	const std::string &	term,
		const std::string &	synonym
	)		const

Remove a synonym for a term.

Parameters

term	The term to remove a synonym for.
synonym	The synonym to remove. If this isn't currently a synonym for term, then no action is taken.

◆ replace_document() [1/2]

Xapian::docid Xapian::WritableDatabase::replace_document	(	const std::string &	unique_term,
		const Xapian::Document &	document
	)

Replace any documents matching a term.

This method replaces any documents indexed by the specified term with the specified document. If any documents are indexed by the term, the lowest document ID will be used for the document, otherwise a new document ID will be generated as for add_document.

One common use is to allow UIDs from another system to easily be mapped to terms in Xapian. Note that this method doesn't automatically add unique_term as a term, so you'll need to call document.add_term(unique_term) first when using replace_document() in this way.

Note that changes to the database won't be immediately committed to disk; see commit() for more details.

As with all database modification operations, the effect is atomic: the document(s) will either be fully replaced, or the document(s) fail to be replaced and an exception is thrown (possibly at a later time when commit() is called or the database is closed).

Parameters

unique_term	The "unique" term.
document	The new document.

Returns: The document ID that document was given.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while writing to the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.

◆ replace_document() [2/2]

void Xapian::WritableDatabase::replace_document	(	Xapian::docid	did,
		const Xapian::Document &	document
	)

Replace a given document in the database.

This method replaces the document with the specified document ID. If document ID did isn't currently used, the document will be added with document ID did.

The monotonic counter used for automatically allocating document IDs is increased so that the next automatically allocated document ID will be did + 1. Be aware that if you use this method to specify a high document ID for a new document, and also use WritableDatabase::add_document(), Xapian may get to a state where this counter wraps around and will be unable to automatically allocate document IDs!

Note that changes to the database won't be immediately committed to disk; see commit() for more details.

As with all database modification operations, the effect is atomic: the document will either be fully replaced, or the document fails to be replaced and an exception is thrown (possibly at a later time when commit() is called or the database is closed).

Parameters

did	The document ID of the document to be replaced.
document	The new document.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while writing to the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.

◆ set_metadata()

void Xapian::WritableDatabase::set_metadata	(	const std::string &	key,
		const std::string &	metadata
	)

Set the user-specified metadata associated with a given key.

This method sets the metadata value associated with a given key. If there is already a metadata value stored in the database with the same key, the old value is replaced. If you want to delete an existing item of metadata, just set its value to the empty string.

User-specified metadata allows you to store arbitrary information in the form of (key, value) pairs.

There's no hard limit on the number of metadata items, or the size of the metadata values. Metadata keys have a limited length, which depend on the backend. We recommend limiting them to 200 bytes. Empty keys are not valid, and specifying one will cause an exception.

Metadata modifications are committed to disk in the same way as modifications to the documents in the database are: i.e., modifications are atomic, and won't be committed to disk immediately (see commit() for more details). This allows metadata to be used to link databases with versioned external resources by storing the appropriate version number in a metadata item.

You can also use the metadata to store arbitrary extra information associated with terms, documents, or postings by encoding the termname and/or document id into the metadata key.

Parameters

key	The key of the metadata item to set.
metadata	The value of the metadata item to set.

Exceptions

Xapian::DatabaseError	will be thrown if a problem occurs while writing to the database.
Xapian::DatabaseCorruptError	will be thrown if the database is in a corrupt state.
Xapian::InvalidArgumentError	will be thrown if the key supplied is empty.
Xapian::UnimplementedError	will be thrown if the database backend in use doesn't support user-specified metadata.

The documentation for this class was generated from the following file:

xapian/database.h

Public Member Functions

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ ~WritableDatabase()

◆ WritableDatabase() [1/3]

◆ WritableDatabase() [2/3]

◆ WritableDatabase() [3/3]

Member Function Documentation

◆ add_database()

◆ add_document()

◆ add_spelling()

◆ add_synonym()

◆ begin_transaction()

◆ cancel_transaction()

◆ clear_synonyms()

◆ commit()

◆ commit_transaction()

◆ delete_document() [1/2]

◆ delete_document() [2/2]

◆ flush()

◆ get_description()

◆ operator=()

◆ remove_spelling()

◆ remove_synonym()

◆ replace_document() [1/2]

◆ replace_document() [2/2]

◆ set_metadata()