WT_CURSOR Struct Reference

A WT_CURSOR handle is the interface to a cursor. More...

Public Member Functions
int	close (WT_CURSOR *cursor)
	Close the cursor. More...
int	largest_key (WT_CURSOR *cursor)
	Get the table's largest key, ignoring visibility. More...
int	reconfigure (WT_CURSOR *cursor, const char *config)
	Reconfigure the cursor. More...
Data access
int	get_key (WT_CURSOR *cursor,...)
	Get the key for the current record. More...
int	get_value (WT_CURSOR *cursor,...)
	Get the value for the current record. More...
void	set_key (WT_CURSOR *cursor,...)
	Set the key for the next operation. More...
void	set_value (WT_CURSOR *cursor,...)
	Set the value for the next operation. More...
Cursor positioning
int	compare (WT_CURSOR *cursor, WT_CURSOR *other, int *comparep)
	Return the ordering relationship between two cursors: both cursors must have the same data source and have valid keys. More...
int	equals (WT_CURSOR *cursor, WT_CURSOR *other, int *equalp)
	Return the ordering relationship between two cursors, testing only for equality: both cursors must have the same data source and have valid keys. More...
int	next (WT_CURSOR *cursor)
	Return the next record. More...
int	prev (WT_CURSOR *cursor)
	Return the previous record. More...
int	reset (WT_CURSOR *cursor)
	Reset the cursor. More...
int	search (WT_CURSOR *cursor)
	Return the record matching the key. More...
int	search_near (WT_CURSOR *cursor, int *exactp)
	Return the record matching the key if it exists, or an adjacent record. More...
Data modification
int	insert (WT_CURSOR *cursor)
	Insert a record and optionally update an existing record. More...
int	modify (WT_CURSOR *cursor, WT_MODIFY *entries, int nentries)
	Modify an existing record. More...
int	update (WT_CURSOR *cursor)
	Update an existing record and optionally insert a record. More...
int	remove (WT_CURSOR *cursor)
	Remove a record. More...
int	reserve (WT_CURSOR *cursor)
	Reserve an existing record so a subsequent write is less likely to fail due to a conflict between concurrent operations. More...

Public Attributes
WT_SESSION *	session
	The session handle for this cursor.
const char *	uri
	The name of the data source for the cursor, matches the uri parameter to WT_SESSION::open_cursor used to open the cursor.
const char *	key_format
	The format of the data packed into key items. More...
const char *	value_format
	The format of the data packed into value items. More...

Detailed Description

A WT_CURSOR handle is the interface to a cursor.

Cursors allow data to be searched, iterated and modified, implementing the CRUD (create, read, update and delete) operations. Cursors are opened in the context of a session. If a transaction is started, cursors operate in the context of the transaction until the transaction is resolved.

Raw data is represented by key/value pairs of WT_ITEM structures, but cursors can also provide access to fields within the key and value if the formats are described in the WT_SESSION::create method.

In the common case, a cursor is used to access records in a table. However, cursors can be used on subsets of tables (such as a single column or a projection of multiple columns), as an interface to statistics, configuration data or application-specific data sources. See WT_SESSION::open_cursor for more information.

Thread safety: A WT_CURSOR handle is not usually shared between threads, see Multithreading for more information.

Examples

ex_access.c, ex_backup.c, ex_backup_block.c, ex_call_center.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_file_system.c, ex_log.c, ex_schema.c, ex_stat.c, and ex_thread.c.

Member Function Documentation

close()

int WT_CURSOR::close

(

WT_CURSOR *

cursor

)

Close the cursor.

This releases the resources associated with the cursor handle. Cursors are closed implicitly by ending the enclosing connection or closing the session in which they were opened.

    error_check(cursor->close(cursor));

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_backup.c, ex_backup_block.c, ex_call_center.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_file_system.c, ex_log.c, ex_schema.c, and ex_stat.c.

compare()

int WT_CURSOR::compare	(	WT_CURSOR *	cursor,
		WT_CURSOR *	other,
		int *	comparep
	)

Return the ordering relationship between two cursors: both cursors must have the same data source and have valid keys.

(When testing only for equality, WT_CURSOR::equals may be faster.)

            int compare;
            error_check(cursor->compare(cursor, other, &compare));
            if (compare == 0) {
                /* Cursors reference the same key */
            } else if (compare < 0) {
                /* Cursor key less than other key */
            } else if (compare > 0) {
                /* Cursor key greater than other key */
            }

Parameters

cursor	the cursor handle
other	another cursor handle
comparep	the status of the comparison: < 0 if cursor refers to a key that appears before other, 0 if the cursors refer to the same key, and > 0 if cursor refers to a key that appears after other.

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

equals()

int WT_CURSOR::equals	(	WT_CURSOR *	cursor,
		WT_CURSOR *	other,
		int *	equalp
	)

Return the ordering relationship between two cursors, testing only for equality: both cursors must have the same data source and have valid keys.

            int equal;
            error_check(cursor->equals(cursor, other, &equal));
            if (equal) {
                /* Cursors reference the same key */
            }

Parameters

	cursor	the cursor handle
	other	another cursor handle
[out]	equalp	the status of the comparison: 1 if the cursors refer to the same key, otherwise 0.

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

get_key()

int WT_CURSOR::get_key	(	WT_CURSOR *	cursor,
			...
	)

Get the key for the current record.

        const char *key; /* Get the cursor's string key. */
        error_check(cursor->get_key(cursor, &key));

        uint64_t recno; /* Get the cursor's record number key. */
        error_check(cursor->get_key(cursor, &recno));

Parameters

cursor	the cursor handle
...	pointers to hold key fields corresponding to WT_CURSOR::key_format. The API does not validate the argument types passed in, the caller is responsible for passing the correct argument types according to WT_CURSOR::key_format.

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_access.c, ex_backup.c, ex_backup_block.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_file_system.c, ex_log.c, ex_schema.c, and ex_thread.c.

get_value()

int WT_CURSOR::get_value	(	WT_CURSOR *	cursor,
			...
	)

Get the value for the current record.

        const char *value; /* Get the cursor's string value. */
        error_check(cursor->get_value(cursor, &value));

        WT_ITEM value; /* Get the cursor's raw value. */
        error_check(cursor->get_value(cursor, &value));

Parameters

cursor	the cursor handle
...	pointers to hold value fields corresponding to WT_CURSOR::value_format. The API does not validate the argument types passed in, the caller is responsible for passing the correct argument types according to WT_CURSOR::value_format.

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_access.c, ex_call_center.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_log.c, ex_schema.c, ex_stat.c, and ex_thread.c.

insert()

int WT_CURSOR::insert

(

WT_CURSOR *

cursor

)

Insert a record and optionally update an existing record.

If the cursor was configured with "overwrite=true" (the default), both the key and value must be set; if the record already exists, the key's value will be updated, otherwise, the record will be inserted.

        /* Insert a new record or overwrite an existing record. */
        const char *key = "some key", *value = "some value";
        error_check(session->open_cursor(session, "table:mytable", NULL, NULL, &cursor));
        cursor->set_key(cursor, key);
        cursor->set_value(cursor, value);
        error_check(cursor->insert(cursor));

If the cursor was not configured with "overwrite=true", both the key and value must be set and the record must not already exist; the record will be inserted. If the record already exists, the WT_DUPLICATE_KEY error is returned and the value found in the tree can be retrieved using WT_CURSOR::get_value.

        /* Insert a new record and fail if the record exists. */
        const char *key = "new key", *value = "some value";
        error_check(
          session->open_cursor(session, "table:mytable", NULL, "overwrite=false", &cursor));
        cursor->set_key(cursor, key);
        cursor->set_value(cursor, value);
        error_check(cursor->insert(cursor));

If a cursor with record number keys was configured with "append=true" (not the default), the value must be set; a new record will be appended and the record number set as the cursor key value.

        /* Insert a new record and assign a record number. */
        uint64_t recno;
        const char *value = "some value";
        cursor->set_value(cursor, value);
        error_check(cursor->insert(cursor));
        error_check(cursor->get_key(cursor, &recno));

The cursor ends with no position, and a subsequent call to the WT_CURSOR::next (WT_CURSOR::prev) method will iterate from the beginning (end) of the table.

If the cursor does not have record number keys or was not configured with "append=true", the cursor ends with no key set and a subsequent call to the WT_CURSOR::get_key method will fail. The cursor ends with no value set and a subsequent call to the WT_CURSOR::get_value method will fail, except for the WT_DUPLICATE_KEY error return, in which case the value currently stored for the key can be retrieved.

Inserting a new record after the current maximum record in a fixed-length bit field column-store (that is, a store with an 'r' type key and 't' type value) may implicitly create the missing records as records with a value of 0.

When loading a large amount of data into a new object, using a cursor with the bulk configuration string enabled and loading the data in sorted order will be much faster than doing out-of-order inserts. See Bulk-load for more information.

The maximum length of a single column stored in a table is not fixed (as it partially depends on the underlying file configuration), but is always a small number of bytes less than 4GB.

The WT_CURSOR::insert method can only be used at snapshot isolation.

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details. In particular, if overwrite=false is configured and a record with the specified key already exists, WT_DUPLICATE_KEY is returned. Also, if in_memory is configured for the database and the insert requires more than the configured cache size to complete, WT_CACHE_FULL is returned.

Examples

ex_access.c, ex_backup.c, ex_backup_block.c, ex_call_center.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_file_system.c, ex_log.c, ex_schema.c, ex_stat.c, and ex_thread.c.

largest_key()

int WT_CURSOR::largest_key

(

WT_CURSOR *

cursor

)

Get the table's largest key, ignoring visibility.

This method is only supported by file: or table: objects. The cursor ends with no position.

        const char *largest_key;
        error_check(cursor->largest_key(cursor));
        error_check(cursor->get_key(cursor, &largest_key));

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

modify()

int WT_CURSOR::modify	(	WT_CURSOR *	cursor,
		WT_MODIFY *	entries,
		int	nentries
	)

Modify an existing record.

Both the key and value must be set and the record must already exist.

Modifications are specified in WT_MODIFY structures. Modifications are applied in order and later modifications can update earlier ones.

The modify method is only supported on strings (value format type S), or raw byte arrays accessed using a WT_ITEM structure (value format type u).

The WT_CURSOR::modify method stores a change record in cache and writes a change record to the log instead of the usual complete values. Using WT_CURSOR::modify will result in slower reads, and slower writes than the WT_CURSOR::insert or WT_CURSOR::update methods, because of the need to assemble the complete value in both the read and write paths. The WT_CURSOR::modify method is intended for applications where memory and log amplification are issues (in other words, applications where there is cache or I/O pressure and the application wants to trade performance for a smaller working set in cache and smaller log records).

        WT_MODIFY entries[3];
        const char *key = "some key";
 
        /* Position the cursor. */
        cursor->set_key(cursor, key);
        error_check(cursor->search(cursor));
 
        /* Replace 20 bytes starting at byte offset 5. */
        entries[0].data.data = "some data";
        entries[0].data.size = strlen(entries[0].data.data);
        entries[0].offset = 5;
        entries[0].size = 20;
 
        /* Insert data at byte offset 40. */
        entries[1].data.data = "and more data";
        entries[1].data.size = strlen(entries[1].data.data);
        entries[1].offset = 40;
        entries[1].size = 0;
 
        /* Replace 2 bytes starting at byte offset 10. */
        entries[2].data.data = "and more data";
        entries[2].data.size = strlen(entries[2].data.data);
        entries[2].offset = 10;
        entries[2].size = 2;
 
        error_check(cursor->modify(cursor, entries, 3));

On success, the cursor ends positioned at the modified record; to minimize cursor resources, the WT_CURSOR::reset method should be called as soon as the cursor no longer needs that position.

The maximum length of a single column stored in a table is not fixed (as it partially depends on the underlying file configuration), but is always a small number of bytes less than 4GB.

The WT_CURSOR::modify method can only be used at snapshot isolation.

Parameters

cursor	the cursor handle
entries	an array of modification data structures
nentries	the number of modification data structures

Returns

zero on success and a non-zero error code on failure. See Error handling for details. In particular, if in_memory is configured for the database and the modify requires more than the configured cache size to complete, WT_CACHE_FULL is returned.

next()

int WT_CURSOR::next

(

WT_CURSOR *

cursor

)

Return the next record.

    error_check(cursor->next(cursor));

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_access.c, ex_backup.c, ex_backup_block.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_file_system.c, ex_log.c, ex_schema.c, ex_stat.c, and ex_thread.c.

prev()

int WT_CURSOR::prev

(

WT_CURSOR *

cursor

)

Return the previous record.

    error_check(cursor->prev(cursor));

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_call_center.c, ex_col_store.c, and ex_cursor.c.

reconfigure()

int WT_CURSOR::reconfigure	(	WT_CURSOR *	cursor,
		const char *	config
	)

Reconfigure the cursor.

The cursor is reset.

        error_check(
          session->open_cursor(session, "table:mytable", NULL, "overwrite=false", &cursor));
 
        /* Reconfigure the cursor to overwrite the record. */
        error_check(cursor->reconfigure(cursor, "overwrite=true"));
 
        cursor->set_key(cursor, key);
        cursor->set_value(cursor, value);
        error_check(cursor->insert(cursor));

Parameters

cursor

the cursor handle

config

configuration string, see Configuration Strings. Permitted values: Name Effect Values append append the value as a new record, creating a new record number key; valid only for cursors with record number keys. a boolean flag; default false. overwrite configures whether the cursor's insert and update methods check the existing state of the record. If overwrite is false, WT_CURSOR::insert fails with WT_DUPLICATE_KEY if the record exists, WT_CURSOR::update fails with WT_NOTFOUND if the record does not exist. a boolean flag; default true.

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

remove()

int WT_CURSOR::remove

(

WT_CURSOR *

cursor

)

Remove a record.

The key must be set; the key's record will be removed if it exists.

        const char *key = "some key";
        error_check(session->open_cursor(session, "table:mytable", NULL, NULL, &cursor));
        cursor->set_key(cursor, key);
        error_check(cursor->remove(cursor));

Any cursor position does not change: if the cursor was positioned before the WT_CURSOR::remove call, the cursor remains positioned at the removed record; to minimize cursor resources, the WT_CURSOR::reset method should be called as soon as the cursor no longer needs that position. If the cursor was not positioned before the WT_CURSOR::remove call, the cursor ends with no position, and a subsequent call to the WT_CURSOR::next (WT_CURSOR::prev) method will iterate from the beginning (end) of the table.

        const char *key = "non-existent key";
        error_check(session->open_cursor(session, "table:mytable", NULL, NULL, &cursor));
        cursor->set_key(cursor, key);
        /* We expect to get a WT_NOTFOUND error if we try to remove a record that does not exist. */
        if ((ret = cursor->remove(cursor)) == WT_NOTFOUND)
            fprintf(stderr, "cursor.remove: key doesn't exist %s\n", wiredtiger_strerror(ret));
        else
            error_check(ret);

Removing a record in a fixed-length bit field column-store (that is, a store with an 'r' type key and 't' type value) is identical to setting the record's value to 0.

The WT_CURSOR::remove method can only be used at snapshot isolation.

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_col_store.c, ex_cursor.c, and ex_extractor.c.

reserve()

int WT_CURSOR::reserve

(

WT_CURSOR *

cursor

)

Reserve an existing record so a subsequent write is less likely to fail due to a conflict between concurrent operations.

The key must first be set and the record must already exist.

Note that reserve works by doing a special update operation that is not logged and does not change the value of the record. This update is aborted when the enclosing transaction ends regardless of whether it commits or rolls back. Given that, reserve can only be used to detect conflicts between transactions that execute concurrently. It cannot detect all logical conflicts between transactions. For that, some update to the record must be committed.

        const char *key = "some key";
        error_check(session->begin_transaction(session, NULL));
        cursor->set_key(cursor, key);
        error_check(cursor->reserve(cursor));
        error_check(session->commit_transaction(session, NULL));

On success, the cursor ends positioned at the specified record; to minimize cursor resources, the WT_CURSOR::reset method should be called as soon as the cursor no longer needs that position.

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

reset()

int WT_CURSOR::reset

(

WT_CURSOR *

cursor

)

Reset the cursor.

Any resources held by the cursor are released, and the cursor's key and position are no longer valid. Subsequent iterations with WT_CURSOR::next will move to the first record, or with WT_CURSOR::prev will move to the last record.

In the case of a statistics cursor, resetting the cursor refreshes the statistics information returned. Resetting a session statistics cursor resets all the session statistics values to zero.

    error_check(cursor->reset(cursor));

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_access.c, ex_cursor.c, and ex_log.c.

search()

int WT_CURSOR::search

(

WT_CURSOR *

cursor

)

Return the record matching the key.

The key must first be set.

        const char *key = "some key";
        cursor->set_key(cursor, key);
        error_check(cursor->search(cursor));

On success, the cursor ends positioned at the returned record; to minimize cursor resources, the WT_CURSOR::reset method should be called as soon as the record has been retrieved and the cursor no longer needs that position.

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_call_center.c, ex_col_store.c, ex_cursor.c, ex_extractor.c, ex_log.c, ex_schema.c, and ex_stat.c.

search_near()

int WT_CURSOR::search_near	(	WT_CURSOR *	cursor,
		int *	exactp
	)

Return the record matching the key if it exists, or an adjacent record.

An adjacent record is either the smallest record larger than the key or the largest record smaller than the key (in other words, a logically adjacent key).

The key must first be set.

An example of a search for an exact or adjacent match:

    cursor->set_key(cursor, key);
    error_check(cursor->search_near(cursor, &exact));
    if (exact == 0) {
        /* an exact match */
    } else if (exact < 0) {
        /* returned smaller key */
    } else if (exact > 0) {
        /* returned larger key */
    }

An example of a forward scan through the table, where all keys greater than or equal to a specified prefix are included in the scan:

    cursor->set_key(cursor, key);
    error_check(cursor->search_near(cursor, &exact));
    if (exact >= 0) {
        /* include first key returned in the scan */
    }
 
    while ((ret = cursor->next(cursor)) == 0) {
        /* the rest of the scan */
    }
    scan_end_check(ret == WT_NOTFOUND);

An example of a backward scan through the table, where all keys less than a specified prefix are included in the scan:

    cursor->set_key(cursor, key);
    error_check(cursor->search_near(cursor, &exact));
    if (exact < 0) {
        /* include first key returned in the scan */
    }
 
    while ((ret = cursor->prev(cursor)) == 0) {
        /* the rest of the scan */
    }
    scan_end_check(ret == WT_NOTFOUND);

On success, the cursor ends positioned at the returned record; to minimize cursor resources, the WT_CURSOR::reset method should be called as soon as the record has been retrieved and the cursor no longer needs that position.

Parameters

cursor	the cursor handle
exactp	the status of the search: 0 if an exact match is found, < 0 if a smaller key is returned, > 0 if a larger key is returned

Returns

zero on success and a non-zero error code on failure. See Error handling for details.

Examples

ex_call_center.c, ex_col_store.c, and ex_cursor.c.

set_key()

void WT_CURSOR::set_key	(	WT_CURSOR *	cursor,
			...
	)

Set the key for the next operation.

        /* Set the cursor's string key. */
        const char *key = "another key";
        cursor->set_key(cursor, key);

        uint64_t recno = 37; /* Set the cursor's record number key. */
        cursor->set_key(cursor, recno);

Parameters

cursor	the cursor handle
...	key fields corresponding to WT_CURSOR::key_format.

If an error occurs during this operation, a flag will be set in the cursor, and the next operation to access the key will fail. This simplifies error handling in applications.

Examples

ex_access.c, ex_backup.c, ex_backup_block.c, ex_call_center.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_file_system.c, ex_log.c, ex_schema.c, ex_stat.c, and ex_thread.c.

set_value()

void WT_CURSOR::set_value	(	WT_CURSOR *	cursor,
			...
	)

Set the value for the next operation.

        /* Set the cursor's string value. */
        const char *value = "another value";
        cursor->set_value(cursor, value);

        WT_ITEM value; /* Set the cursor's raw value. */
        value.data = "another value";
        value.size = strlen("another value");
        cursor->set_value(cursor, &value);

Parameters

cursor	the cursor handle
...	value fields corresponding to WT_CURSOR::value_format.

If an error occurs during this operation, a flag will be set in the cursor, and the next operation to access the value will fail. This simplifies error handling in applications.

Examples

ex_access.c, ex_backup.c, ex_backup_block.c, ex_call_center.c, ex_col_store.c, ex_cursor.c, ex_encrypt.c, ex_extractor.c, ex_file_system.c, ex_log.c, ex_schema.c, ex_stat.c, and ex_thread.c.

update()

int WT_CURSOR::update

(

WT_CURSOR *

cursor

)

Update an existing record and optionally insert a record.

If the cursor was configured with "overwrite=true" (the default), both the key and value must be set; if the record already exists, the key's value will be updated, otherwise, the record will be inserted.

        const char *key = "some key", *value = "some value";
        error_check(session->open_cursor(session, "table:mytable", NULL, NULL, &cursor));
        cursor->set_key(cursor, key);
        cursor->set_value(cursor, value);
        error_check(cursor->update(cursor));

If the cursor was not configured with "overwrite=true", both the key and value must be set and the record must already exist; the record will be updated.

        const char *key = "some key", *value = "some value";
        error_check(
          session->open_cursor(session, "table:mytable", NULL, "overwrite=false", &cursor));
        cursor->set_key(cursor, key);
        cursor->set_value(cursor, value);
        error_check(cursor->update(cursor));

On success, the cursor ends positioned at the modified record; to minimize cursor resources, the WT_CURSOR::reset method should be called as soon as the cursor no longer needs that position. (The WT_CURSOR::insert method never keeps a cursor position and may be more efficient for that reason.)

The maximum length of a single column stored in a table is not fixed (as it partially depends on the underlying file configuration), but is always a small number of bytes less than 4GB.

The WT_CURSOR::update method can only be used at snapshot isolation.

Parameters

cursor

the cursor handle

Returns

zero on success and a non-zero error code on failure. See Error handling for details. In particular, if overwrite=false is configured and no record with the specified key exists, WT_NOTFOUND is returned. Also, if in_memory is configured for the database and the update requires more than the configured cache size to complete, WT_CACHE_FULL is returned.

Examples

ex_col_store.c, ex_cursor.c, and ex_schema.c.

Member Data Documentation

key_format

const char* WT_CURSOR::key_format

The format of the data packed into key items.

See Packing and Unpacking Data for details. If not set, a default value of "u" is assumed, and applications must use WT_ITEM structures to manipulate untyped byte arrays.

value_format

const char* WT_CURSOR::value_format

The format of the data packed into value items.

See Packing and Unpacking Data for details. If not set, a default value of "u" is assumed, and applications must use WT_ITEM structures to manipulate untyped byte arrays.