@silverstripe.com)
*/
abstract class BulkLoader extends ViewableData {
/**
* Each row in the imported dataset should map to one instance
* of this class (with optional property translation
* through {@self::$columnMaps}.
*
* @var string
*/
public $objectClass;
/**
* Override this on subclasses to give the specific functions names.
*
* @var string
*/
public static $title;
/**
* Map columns to DataObject-properties.
* If not specified, we assume the first row
* in the file contains the column headers.
* The order of your array should match the column order.
*
* The column count should match the count of array elements,
* fill with NULL values if you want to skip certain columns.
*
* You can also combine {@link $hasHeaderRow} = true and {@link $columnMap}
* and omit the NULL values in your map.
*
* Supports one-level chaining of has_one relations and properties with dot notation
* (e.g. Team.Title). The first part has to match a has_one relation name
* (not necessarily the classname of the used relation).
*
*
* 'FirstName', // custom column name
* null, // ignored column
* 'RegionID', // direct has_one/has_many ID setting
* 'OrganisationTitle', // create has_one relation to existing record using $relationCallbacks
* 'street' => 'Organisation.StreetName', // match an existing has_one or create one and write property.
* );
* ?>
*
*
* @var array
*/
public $columnMap = array();
/**
* Find a has_one relation based on a specific column value.
*
*
* array(
* 'relationname' => 'Organisation', // relation accessor name
* 'callback' => 'getOrganisationByTitle',
* );
* );
* ?>
*
*
* @var array
*/
public $relationCallbacks = array();
/**
* Specifies how to determine duplicates based on one or more provided fields
* in the imported data, matching to properties on the used {@link DataObject} class.
* Alternatively the array values can contain a callback method (see example for
* implementation details).
* If multiple checks are specified, the first one "wins".
*
*
* 'ID',
* 'phonenumber' => array(
* 'callback' => 'getByImportedPhoneNumber'
* )
* );
* ?>
*
*
* @var array
*/
public $duplicateChecks = array();
function __construct($objectClass) {
$this->objectClass = $objectClass;
ini_set('max_execution_time', 3600);
ini_set('memory_limit', '512M');
}
/*
* Load the given file via {@link self::processAll()} and {@link self::processRecord()}.
*
* @return array See {@link self::processAll()}
*/
public function load($filepath) {
return $this->processAll($filepath);
}
/**
* Preview a file import (don't write anything to the database).
* Useful to analyze the input and give the users a chance to influence
* it through a UI.
*
* @todo Implement preview()
*
* @param string $filepath Absolute path to the file we're importing
* @return array See {@link self::processAll()}
*/
public function preview($filepath) {
user_error("BulkLoader::preview(): Not implemented", E_USER_ERROR);
}
/**
* Process every record in the file
*
* @param string $filepath Absolute path to the file we're importing (with UTF8 content)
* @param boolean $preview If true, we'll just output a summary of changes but not actually do anything
* @return array Information about the import process, with each row matching a created or updated DataObject.
* Array structure:
* - 'id': Database id of the created or updated record
* - 'action': Performed action ('create', 'update')
* - 'message': free-text string that can optionally provide some more information about what changes have
*/
abstract protected function processAll($filepath, $preview = false);
/**
* Process a single record from the file.
*
* @param array $record An map of the data, keyed by the header field defined in {@link self::$columnMap}
* @param array $columnMap
* @param boolean $preview
* @return ArrayData @see self::processAll()
*/
abstract protected function processRecord($record, $columnMap, $preview = false);
/**
* Return a FieldSet containing all the options for this form; this
* doesn't include the actual upload field itself
*/
public function getOptionFields() {}
/**
* Return a human-readable name for this object.
* It defaults to the class name can be overridden by setting the static variable $title
*
* @return string
*/
public function Title() {
return ($title = $this->stat('title')) ? $title : $this->class;
}
/**
* Get a specification of all available columns and relations on the used model.
* Useful for generation of spec documents for technical end users.
*
* Return Format:
*
* array(
* 'fields' => array('myFieldName'=>'myDescription'),
* 'relations' => array('myRelationName'=>'myDescription'),
* )
*
*
* @todo Mix in custom column mappings
* @usedby {@link ModelAdmin}
*
* @return array
**/
public function getImportSpec() {
$spec = array();
// get database columns (fieldlabels include fieldname as a key)
$spec['fields'] = (array)singleton($this->objectClass)->fieldLabels();
$has_ones = singleton($this->objectClass)->has_one();
$has_manys = singleton($this->objectClass)->has_many();
$many_manys = singleton($this->objectClass)->many_many();
$spec['relations'] = (array)$has_ones + (array)$has_manys + (array)$many_manys;
return $spec;
}
/**
* Determines if a specific field is null.
* Can be useful for unusual "empty" flags in the file,
* e.g. a "(not set)" value.
* The usual {@link DBField::isNull()} checks apply when writing the {@link DataObject},
* so this is mainly a customization method.
*
* @param mixed $val
* @param string $field Name of the field as specified in the array-values for {@link self::$columnMap}.
* @return boolean
*/
protected function isNullValue($val, $fieldName = null) {
return (empty($val));
}
}
?>