Difference between revisions of "API:EPrints/Database"

From EPrints Documentation
Jump to: navigation, search
Line 573: Line 573:
  
 
  $str = $db->quote_binary( $bytes )
 
  $str = $db->quote_binary( $bytes )
Some databases (PostgreSQL) require weird transforms of binary data to work correctly.
+
Some databases (Oracle/PostgreSQL) require transforms of binary data to work correctly.
  
 
This method should be called on data containing nul bytes or back-slashes before being passed on {{API:PodLink|file=|package_name=|section=quote_value|text=/quote_value}}.
 
This method should be called on data containing nul bytes or back-slashes before being passed on {{API:PodLink|file=|package_name=|section=quote_value|text=/quote_value}}.
 +
 +
<div style='background-color: #e8e8f; margin: 0.5em 0em 1em 0em; border: solid 1px #cce;  padding: 0em 1em 0em 1em; font-size: 80%; '>
 +
<span style='display:none'>User Comments</span>
 +
<!-- Edit below this comment -->
 +
 +
 +
<!-- Pod2Wiki= -->
 +
</div>
 +
<!-- Pod2Wiki=item_quote_ordervalue -->
 +
===quote_ordervalue===
 +
 +
$str = $db-&gt;quote_ordervalue( $field, $value )
 +
Some databases (Oracle) can't order by CLOBS so need special treatment when creating the ordervalues tables. This method allows any fixing-up required for string data before it's inserted.
  
 
<div style='background-color: #e8e8f; margin: 0.5em 0em 1em 0em; border: solid 1px #cce;  padding: 0em 1em 0em 1em; font-size: 80%; '>
 
<div style='background-color: #e8e8f; margin: 0.5em 0em 1em 0em; border: solid 1px #cce;  padding: 0em 1em 0em 1em; font-size: 80%; '>
Line 1,401: Line 1,414:
 
<!-- Pod2Wiki= -->
 
<!-- Pod2Wiki= -->
 
</div>
 
</div>
<!-- Pod2Wiki=item_alias_glue -->
+
<!-- Pod2Wiki=item_sql_as -->
===alias_glue===
+
===sql_as===
  
  $glue = $db-&gt;alias_glue()
+
  $sql = $db-&gt;sql_AS()
 
Returns the syntactic glue to use when aliasing. SQL 92 DBs will happilly use " AS " but some DBs (Oracle!) won't accept it.
 
Returns the syntactic glue to use when aliasing. SQL 92 DBs will happilly use " AS " but some DBs (Oracle!) won't accept it.
 +
 +
<div style='background-color: #e8e8f; margin: 0.5em 0em 1em 0em; border: solid 1px #cce;  padding: 0em 1em 0em 1em; font-size: 80%; '>
 +
<span style='display:none'>User Comments</span>
 +
<!-- Edit below this comment -->
 +
 +
 +
<!-- Pod2Wiki= -->
 +
</div>
 +
<!-- Pod2Wiki=item_sql_like -->
 +
===sql_like===
 +
 +
$sql = $db-&gt;sql_LIKE()
 +
Returns the syntactic glue to use when making a case-insensitive LIKE. PostgreSQL requires "ILIKE" while everything else uses "LIKE" and the column collation.
  
 
<div style='background-color: #e8e8f; margin: 0.5em 0em 1em 0em; border: solid 1px #cce;  padding: 0em 1em 0em 1em; font-size: 80%; '>
 
<div style='background-color: #e8e8f; margin: 0.5em 0em 1em 0em; border: solid 1px #cce;  padding: 0em 1em 0em 1em; font-size: 80%; '>

Revision as of 13:05, 17 June 2010

EPrints 3 Reference: Directory Structure - Metadata Fields - Repository Configuration - XML Config Files - XML Export Format - EPrints data structure - Core API - Data Objects


API: Core API

Latest Source Code (3.4, 3.3) | Revision Log | Before editing this page please read Pod2Wiki


Contents

NAME

EPrints::Database - a connection to the SQL database for an eprints session.

User Comments


DESCRIPTION

EPrints Database Access Module

Provides access to the backend database. All database access done via this module, in the hope that the backend can be replaced as easily as possible.

The database object is created automatically when you start a new eprints session. To get a handle on it use:

$db = $session->get_database

User Comments


Cross-database Support

Any use of SQL must use quote_identifier to quote database tables and columns. The only exception to this are the Database::* modules which provide database-driver specific extensions.

Variables that are database quoted are prefixed with 'Q_'.

User Comments


METHODS

User Comments


build_connection_string

$dbstr = EPrints::Database::build_connection_string( %params )

Build the string to use to connect to the database via DBI. %params must contain dbname, and may also contain dbport, dbhost and dbsock.

User Comments


create

$db = $db->create( $username, $password )

Create and connect to a new database using super user account $username and $password.

User Comments


new

$db = EPrints::Database->new( $session )

Create a connection to the database.

User Comments


connect

$foo = $db->connect

Connects to the database.

User Comments


disconnect

$foo = $db->disconnect

Disconnects from the EPrints database. Should always be done before any script exits.

User Comments


error

$errstr = $db->error

Return a string describing the last SQL error.

User Comments


begin

$db->begin

Begin a transaction.

User Comments


commit

$db->commit

Commit the previous begun transaction.

User Comments


rollback

$db->rollback

Rollback the partially completed transaction.

User Comments


create_archive_tables

$success = $db->create_archive_tables

Create all the SQL tables for each dataset.

User Comments


drop_archive_tables

$db->drop_archive_tables()

Destroy all tables used by eprints in the database.

User Comments


create_dataset_tables

$success = $db->create_dataset_tables( $dataset )

Create all the SQL tables for a single dataset.

User Comments


drop_dataset_tables

$db->drop_dataset_tables( $dataset )

Drop all the SQL tables for a single dataset.

User Comments


create_dataset_index_tables

$success = $db->create_dataset_index_tables( $dataset )

Create all the index tables for a single dataset.

User Comments


create_dataset_ordervalues_tables

$success = $db->create_dataset_ordervalues_tables( $dataset )

Create all the ordervalues tables for a single dataset.

User Comments


type_info

$type_info = $db->type_info( DATA_TYPE )

See DBI/type_info.

User Comments


get_column_type

$real_type = $db->get_column_type( NAME, TYPE, NOT_NULL, [ LENGTH/PRECISION ], [ SCALE ], %opts )

Returns a SQL column definition for NAME of type TYPE, usually something like:

 $name $type($length,$scale) [ NOT NULL ]
 

If NOT_NULL is true column will be set "not null".

LENGTH/PRECISION and SCALE control the maximum lengths of character or decimal types (see below).

Other options available to refine the column definition:

 langid - character set/collation to use
 sorted - whether this column will be used to order by
 

langid is mapped to real database values by the "dblanguages" configuration option. The database may not be able to order the request column type in which case, if sorted is true, the database may use a substitute column type.

TYPE is the SQL type. The types are constants defined by this module, to import them use:

 use EPrints::Database qw( :sql_types );
 

Supported types (n = requires LENGTH argument):

Character data: SQL_VARCHAR(n), SQL_LONGVARCHAR.

Binary data: SQL_VARBINARY(n), SQL_LONGVARBINARY.

Integer data: SQL_TINYINT, SQL_SMALLINT, SQL_INTEGER, SQL_BIGINT.

Floating-point data: SQL_REAL, SQL_DOUBLE.

Time data: SQL_DATE, SQL_TIME.

The actual column types used will be database-specific.

User Comments


create_table

$success = $db->create_table( $tablename, $dataset, $setkey, @fields );

Create the tables used to store metadata for this dataset: the main table and any required for multiple or mulitlang fields.

The first $setkey number of fields are used for a primary key.

User Comments


has_sequence

$boolean = $db->has_sequence( $name )

Return true if a sequence of the given name exists in the database.

User Comments


create_sequence

$success = $db->create_sequence( $seq_name )

Creates a new sequence object initialised to zero.

User Comments


drop_sequence

$success = $db->drop_sequence( $seq_name )

Deletes a sequence object.

User Comments


drop_column

$success = $db->drop_column( $table, $column )

Drops a column from a table.

User Comments


get_primary_key

@columns = $db->get_primary_key( $tablename )

Returns the list of column names that comprise the primary key for $tablename.

Returns empty list if no primary key exists.

User Comments


create_index

$success = $db->create_index( $tablename, @columns )

Creates an index over @columns for $tablename. Returns true on success.

User Comments


create_unique_index

$success = $db->create_unique_index( $tablename, @columns )

Creates a unique index over @columns for $tablename. Returns true on success.

User Comments


_update

$success = $db->_update( $tablename, $keycols, $keyvals, $columns, @values )

UPDATES $tablename where $keycols equals $keyvals.

This method is internal.

User Comments


_update_quoted

$success = $db->_update_quoted( $tablename, $keycols, $keyvals, $columns, @qvalues )

UPDATES $tablename where $keycols equals $keyvals. Won't quote $keyvals or @qvalues before use - use this method with care!

This method is internal.

User Comments


insert

$success = $db->insert( $table, $columns, @values )

Inserts values into the table $table. If $columns is defined it will be used as a list of columns to insert into. @values is a list of arrays containing values to insert.

Values will be quoted before insertion.

User Comments


insert_quoted

$success = $db->insert_quoted( $table, $columns, @qvalues )

Inserts values into the table $table. If $columns is defined it will be used as a list of columns to insert into. @qvalues is a list of arrays containing values to insert.

Values will NOT be quoted before insertion - care must be exercised!

User Comments


delete_from

$success = $db->delete_from( $table, $columns, @values )

Perform a SQL DELETE FROM $table using $columns to build a where clause. @values is a list of array references of values in the same order as $columns.

If you want to clear a table completely use clear_table().

User Comments


add_record

$success = $db->add_record( $dataset, $data )

Add the given data as a new record in the given dataset. $data is a reference to a hash containing values structured for a record in the that dataset.

User Comments


prep_int

$mungedvalue = EPrints::Database::prep_int( $value )

Escape a numerical value for SQL. undef becomes NULL. Anything else becomes a number (zero if needed).

User Comments


prep_value

$mungedvalue = EPrints::Database::prep_value( $value )

Escape a value for SQL. Modify value such that " becomes \" and \ becomes \\ and ' becomes \'

User Comments


prep_like_value

$mungedvalue = EPrints::Database::prep_like_value( $value )

Escape an value for an SQL like field. In addition to ' " and \ also escapes % and _

User Comments


quote_value

$str = $db->quote_value( $value )

Return a quoted value. To quote a 'like' value you should do:

my $str = $database->quote_value( EPrints::Database::prep_like_value( $foo ) . '%' );
 

User Comments


quote_int

$str = $db->quote_int( $value )

Return a quoted integer value

User Comments


quote_binary

$str = $db->quote_binary( $bytes )

Some databases (Oracle/PostgreSQL) require transforms of binary data to work correctly.

This method should be called on data containing nul bytes or back-slashes before being passed on /quote_value.

User Comments


quote_ordervalue

$str = $db->quote_ordervalue( $field, $value )

Some databases (Oracle) can't order by CLOBS so need special treatment when creating the ordervalues tables. This method allows any fixing-up required for string data before it's inserted.

User Comments


quote_identifier

$str = $db->quote_identifier( @parts )

Quote a database identifier (e.g. table names). Multiple @parts will be joined by dot.

User Comments


update

$success = $db->update( $dataset, $data, $changed, $insert )

Updates a record in the database with the given $data. Obviously the value of the primary key must be set.

This also updates the text indexes and the ordering keys.

User Comments


remove

$success = $db->remove( $dataset, $id )

Attempts to remove the record with the primary key $id from the specified dataset.

User Comments


create_counters

$success = $db->create_counters

Create the counters used to store the highest current id of eprints, users etc.

User Comments


has_counter

$success = $db->has_counter( $counter )

Returns true if $counter exists.

User Comments


create_counter

$success = $db->create_counter( $name )

Create and initialise to zero a new counter called $name.

User Comments


remove_counters

$success = $db->remove_counters

Destroy all counters.

User Comments


drop_counter

$success = $db->drop_counter( $name )

Destroy the counter named $name.

User Comments


next_doc_pos

$n = $db->next_doc_pos( $eprintid )

Return the next unused document pos for the given eprintid.

User Comments


counter_current

$n = $db->counter_current( $counter )

Return the value of the previous counter_next on $counter.

User Comments


counter_next

$n = $db->counter_next( $counter )

Return the next unused value for the named counter. Returns undef if the counter doesn't exist.

User Comments


counter_minimum

$db->counter_minimum( $counter, $value )

Ensure that the counter is set no lower than $value. This is used when importing eprints which may not be in scrict sequence.

User Comments


counter_reset

$db->counter_reset( $counter )

Reset the counter. Use with caution.

User Comments


cache_exp

$searchexp = $db->cache_exp( $cacheid )

Return the serialised Search of a the cached search with id $cacheid. Return undef if the id is invalid or expired.

User Comments


cache

$cacheid = $db->cache( $searchexp, $dataset, $srctable, [$order], [$list] )

Create a cache of the specified search expression from the SQL table $srctable.

If $order is set then the cache is ordered by the specified fields. For example "-year/title" orders by year (descending). Records with the same year are ordered by title.

If $srctable is set to "LIST" then order is ignored and the list of ids is taken from the array reference $list.

If $srctable is set to "ALL" every matching record from $dataset is added to the cache, optionally ordered by $order.

User Comments


cache_table

$tablename = $db->cache_table( $id )

Return the SQL table used to store the cache with id $id.

User Comments


get_index_ids

$ids = $db->get_index_ids( $table, $condition )

Return a reference to an array of the distinct primary keys from the given SQL table which match the specified condition.

User Comments


search

$ids = $db->search( $keyfield, $tables, $conditions, [$main_table_alias] )

Return a reference to an array of ids - the results of the search specified by $conditions accross the tables specified in the $tables hash where keys are tables aliases and values are table names.

If no table alias is passed then M is assumed.

User Comments


drop_cache

$db->drop_cache( $id )

Remove the cached search with the given id.

User Comments


count_table

$n = $db->count_table( $tablename )

Return the number of rows in the specified SQL table.

User Comments


from_cache

$foo = $db->from_cache( $dataset, $cacheid, [$offset], [$count], [$justids] )

Return a reference to an array containing all the items from the given dataset that have id's in the specified cache. The cache may be specified either by id or serialised search expression.

$offset is an offset from the start of the cache and $count is the number of records to return.

If $justids is true then it returns just an ref to an array of the record ids, not the objects.

User Comments


drop_orphan_cache_tables

$c = $db->drop_orphan_cache_tables

Drop tables called "cacheXXX" where XXX is an integer. Returns the number of tables dropped.

User Comments


get_single

$obj = $db->get_single( $dataset, $id )

Return a single item from the given dataset. The one with the specified id.

User Comments


get_all

$items = $db->get_all( $dataset )

Returns a reference to an array with all the items from the given dataset.

User Comments


get_cache_ids

@ids = $db->get_cache_ids( $dataset, $cachemap, $offset, $count )

Returns a list of $count ids from $cache_id starting at $offset and in the order in the cachemap.

User Comments


get_dataobjs

@dataobjs = $db->get_dataobjs( $dataset [, $id [, $id ] ] )

Retrieves the records in $dataset with the given $id(s). If an $id doesn't exist in the database it will be ignored.

User Comments


get_values

$foo = $db->get_values( $field, $dataset )

Return a reference to an array of all the distinct values of the EPrints::MetaField specified.

User Comments


sort_values

$values = $db->sort_values( $field, $values [, $langid ] )

ALPHA!!! Liable to API change!!!

Sorts and returns the list of $values using the database.

$field is used to get the order value for each value. $langid (or $session->get_langid if unset) is used to determine the database collation to use when sorting the resulting order values.

User Comments


get_ids_by_field_values

$ids = $db->get_ids_by_field_values( $field, $dataset [ %opts ] )

Return a reference to a hash table where the keys are field value ids and the value is a reference to an array of ids.

User Comments


do

$success = $db->do( $sql )

Execute the given SQL.

User Comments


prepare

$sth = $db->prepare( $sql )

Prepare the given $sql and return a handle on it.

User Comments


prepare_select

$sth = $db->prepare_select( $sql [, %options ] )

Prepare a SELECT statement $sql and return a handle to it. After preparing a statement use execute() to execute it.

The LIMIT SQL keyword is not universally supported, to specify a LIMIT you must use the limit option.

Options:

 limit - limit the number of rows returned
 offset - return limit number of rows after offset
 

User Comments


execute

$success = $db->execute( $sth, $sql )

Execute the SQL prepared earlier. $sql is only passed in for debugging purposes.

User Comments


has_dataset

$db->has_dataset( $dataset )

Returns true if $dataset exists in the database or has no database tables.

This does not check that all fields are configured - see has_field().

User Comments


has_field

$db->has_field( $dataset, $field )

Returns true if $field is in the database for $dataset.

User Comments


add_field

$db->add_field( $dataset, $field [, $force ] )

Add $field to $dataset's tables.

If $force is true will modify/replace an existing column (use with care!).

User Comments


remove_field

$db->remove_field( $dataset, $field )

Remove $field from $dataset's tables.

User Comments


rename_field

$ok = $db->rename_field( $dataset, $field, $old_name )

Rename a $field in the database from it's old name $old_name.

Returns true if the field was successfully renamed.

User Comments


exists

$boolean = $db->exists( $dataset, $id )

Return true if a record with the given primary key exists in the dataset, otherwise false.

User Comments


set_debug

$db->set_debug( $boolean )

Set the SQL debug mode to true or false.

User Comments


create_version_table

$db->create_version_table

Make the version table (and set the only value to be the current version of eprints).

User Comments


drop_version_table

$db->drop_version_table

Drop the version table.

User Comments


set_version

$db->set_version( $versionid );

Set the version id table in the SQL database to the given value (used by the upgrade script).

User Comments


has_table

$boolean = $db->has_table( $tablename )

Return true if a table of the given name exists in the database.

User Comments


has_column

$boolean = $db->has_column( $tablename, $columnname )

Return true if the a table of the given name has a column named $columnname in the database.

User Comments


index_name

$name = $db->index_name( $table, @columns )

Returns the name of the first index that starts with @columns on the $table table.

Returns undef if no index exists.

User Comments


drop_table

$db->drop_table( $tablename )

Delete the named table. Use with caution!

User Comments


clear_table

$db->clear_table( $tablename )

Clears all records from the given table, use with caution!

User Comments


rename_table

$db->rename_table( $tablename, $newtablename )

Renames the table from the old name to the new one.

User Comments


swap_table

$db->swap_table( $table_a, $table_b )

Swap table a and table b.

User Comments


get_tables

@tables = $db->get_tables

Return a list of all the tables in the database.

User Comments


get_version

$version = $db->get_version

Return the version of eprints which the database is compatable with or undef if unknown (before v2.1).

User Comments


is_latest_version

$boolean = $db->is_latest_version

Return true if the SQL tables are in the correct configuration for this edition of eprints. Otherwise false.

User Comments


valid_login

$db->valid_login( $username, $password )

Returns whether the clear-text $password matches the stored crypted password for $username.

User Comments


get_server_version

$version = $db->get_server_version

Return the database server version.

User Comments


get_default_charset

$charset = $db->get_default_charset( LANGUAGE )

Return the character set to use for LANGUAGE.

Returns undef if character sets are unsupported.

User Comments


get_default_collation

$collation = $db->get_default_collation( LANGUAGE )

Return the collation to use for LANGUAGE.

Returns undef if collation is unsupported.

User Comments


get_driver_name

$driver = $db->get_driver_name

Return the database driver name.

User Comments


dequeue_events

@events = $db->dequeue_events( $n )

Attempt to dequeue upto $n events. May return between 0 and $n events depending on parallel processes and how many events are remaining on the queue.

User Comments


prepare_regexp

$sql = $db->prepare_regexp( $quoted_column, $quoted_value )

The syntax used for regular expressions varies across databases. This method takes two quoted values and returns a SQL expression that will apply the regexp ($quoted_value) to the column ($quoted_column).

User Comments


sql_as

$sql = $db->sql_AS()

Returns the syntactic glue to use when aliasing. SQL 92 DBs will happilly use " AS " but some DBs (Oracle!) won't accept it.

User Comments


sql_like

$sql = $db->sql_LIKE()

Returns the syntactic glue to use when making a case-insensitive LIKE. PostgreSQL requires "ILIKE" while everything else uses "LIKE" and the column collation.

User Comments