DataTable.php

<?php
/**
 * Piwik - Open source web analytics
 *
 * @link http://piwik.org
 * @license http://www.gnu.org/licenses/gpl-3.0.html GPL v3 or later
 *
 * @category Piwik
 * @package Piwik
 */

namespace Piwik;

use Closure;
use Exception;
use Piwik\Common;
use Piwik\DataTable\Manager;
use Piwik\DataTable\Renderer\Html;
use Piwik\DataTable\Row;
use Piwik\DataTable\Filter;
use Piwik\DataTable\Row\DataTableSummaryRow;
use ReflectionClass;

/**
 * @see Common::destroy()
 */
require_once PIWIK_INCLUDE_PATH . '/core/Common.php';

/**
 *
 * ---- DataTable
 * A DataTable is a data structure used to store complex tables of data.
 *
 * A DataTable is composed of multiple DataTable\Row.
 * A DataTable can be applied one or several DataTable_Filter.
 * A DataTable can be given to a DataTable_Renderer that would export the data under a given format (XML, HTML, etc.).
 *
 * A DataTable has the following features:
 * - serializable to be stored in the DB
 * - loadable from the serialized version
 * - efficient way of loading data from an external source (from a PHP array structure)
 * - very simple interface to get data from the table
 *
 * ---- DataTable\Row
 * A DataTableRow in the table is defined by
 * - multiple columns (a label, multiple values, ...)
 * - optional metadata
 * - optional - a sub DataTable associated to this row
 *
 * Simple row example:
 * - columns = array(   'label' => 'Firefox',
 *                        'visitors' => 155,
 *                        'pages' => 214,
 *                        'bounce_rate' => 67)
 * - metadata = array('logo' => '/plugins/UserSettings/images/browsers/FF.gif')
 * - no sub DataTable
 *
 * A more complex example would be a DataTable\Row that is associated to a sub DataTable.
 * For example, for the row of the search engine Google,
 * we want to get the list of keywords associated, with their statistics.
 * - columns = array(   'label' => 'Google',
 *                        'visits' => 1550,
 *                        'visits_length' => 514214,
 *                        'returning_visits' => 77)
 * - metadata = array(    'logo' => '/plugins/Referers/images/searchEngines/google.com.png',
 *                        'url' => 'http://google.com')
 * - DataTable = DataTable containing several DataTable\Row containing the keywords information for this search engine
 *            Example of one DataTable\Row
 *            - the keyword columns specific to this search engine =
 *                    array(  'label' => 'Piwik', // the keyword
 *                            'visitors' => 155,  // Piwik has been searched on Google by 155 visitors
 *                            'pages' => 214 // Visitors coming from Google with the kwd Piwik have seen 214 pages
 *                    )
 *            - the keyword metadata = array() // nothing here, but we could imagining storing the URL of the search in Google for example
 *            - no subTable
 *
 *
 * ---- DataTable_Filter
 * A DataTable_Filter is a applied to a DataTable and so
 * can filter information in the multiple DataTable\Row.
 *
 * For example a DataTable_Filter can:
 * - remove rows from the table,
 *        for example the rows' labels that do not match a given searched pattern
 *        for example the rows' values that are less than a given percentage (low population)
 * - return a subset of the DataTable
 *        for example a function that apply a limit: $offset, $limit
 * - add / remove columns
 *        for example adding a column that gives the percentage of a given value
 * - add some metadata
 *        for example the 'logo' path if the filter detects the logo
 * - edit the value, the label
 * - change the rows order
 *        for example if we want to sort by Label alphabetical order, or by any column value
 *
 * When several DataTable_Filter are to be applied to a DataTable they are applied sequentially.
 * A DataTable_Filter is assigned a priority.
 * For example, filters that
 *    - sort rows should be applied with the highest priority
 *    - remove rows should be applied with a high priority as they prune the data and improve performance.
 *
 * ---- Code example
 *
 * $table = new DataTable();
 * $table->addRowsFromArray( array(...) );
 *
 * # sort the table by visits asc
 * $filter = new DataTable_Filter_Sort( $table, 'visits', 'asc');
 * $tableFiltered = $filter->getTableFiltered();
 *
 * # add a filter to select only the website with a label matching '*.com' (regular expression)
 * $filter = new DataTable_Filter_Pattern( $table, 'label', '*(.com)');
 * $tableFiltered = $filter->getTableFiltered();
 *
 * # keep the 20 elements from offset 15
 * $filter = new DataTable_Filter_Limit( $tableFiltered, 15, 20);
 * $tableFiltered = $filter->getTableFiltered();
 *
 * # add a column computing the percentage of visits
 * # params = table, column containing the value, new column name to add, number of total visits to use to compute the %
 * $filter = new DataTable_Filter_AddColumnPercentage( $tableFiltered, 'visits', 'visits_percentage', 2042);
 * $tableFiltered = $filter->getTableFiltered();
 *
 * # we get the table as XML
 * $xmlOutput = new DataTable_Exporter_Xml( $table );
 * $xmlOutput->setHeader( ... );
 * $xmlOutput->setColumnsToExport( array('visits', 'visits_percent', 'label') );
 * $XMLstring = $xmlOutput->getOutput();
 *
 *
 * ---- Other (ideas)
 * We can also imagine building a DataTable_Compare which would take N DataTable that have the same
 * structure and would compare them, by computing the percentages of differences, etc.
 *
 * For example
 * DataTable1 = [ keyword1, 1550 visits]
 *                [ keyword2, 154 visits ]
 * DataTable2 = [ keyword1, 1004 visits ]
 *                [ keyword3, 659 visits ]
 * DataTable_Compare = result of comparison of table1 with table2
 *                        [ keyword1, +154% ]
 *                        [ keyword2, +1000% ]
 *                        [ keyword3, -430% ]
 *
 * @package Piwik
 * @subpackage DataTable
 */
class DataTable
{
    /** Name for metadata that describes when a report was archived. */
    const ARCHIVED_DATE_METADATA_NAME = 'archived_date';
    const MAX_DEPTH_DEFAULT = 15;
    /** Name for metadata that describes which columns are empty and should not be shown. */
    const EMPTY_COLUMNS_METADATA_NAME = 'empty_column';

    /**
     * Maximum nesting level.
     */
    private static $maximumDepthLevelAllowed = self::MAX_DEPTH_DEFAULT;

    /**
     * Array of Row
     *
     * @var Row[]
     */
    protected $rows = array();

    /**
     * Id assigned to the DataTable, used to lookup the table using the DataTable_Manager
     *
     * @var int
     */
    protected $currentId;

    /**
     * Current depth level of this data table
     * 0 is the parent data table
     *
     * @var int
     */
    protected $depthLevel = 0;

    /**
     * This flag is set to false once we modify the table in a way that outdates the index
     *
     * @var bool
     */
    protected $indexNotUpToDate = true;

    /**
     * This flag sets the index to be rebuild whenever a new row is added,
     * as opposed to re-building the full index when getRowFromLabel is called.
     * This is to optimize and not rebuild the full Index in the case where we
     * add row, getRowFromLabel, addRow, getRowFromLabel thousands of times.
     *
     * @var bool
     */
    protected $rebuildIndexContinuously = false;

    /**