<?php

/**
 * @file
 * Install, update and uninstall functions for the taxonomy module.
 */

/**
 * Implements hook_uninstall().
 */
function taxonomy_uninstall() {
  // Remove variables.
  variable_del('taxonomy_override_selector');
  variable_del('taxonomy_terms_per_page_admin');
  // Remove taxonomy_term bundles.
  $vocabularies = db_query("SELECT machine_name FROM {taxonomy_vocabulary}")->fetchCol();
  foreach ($vocabularies as $vocabulary) {
    field_attach_delete_bundle('taxonomy_term', $vocabulary);
  }
}

/**
 * Implements hook_schema().
 */
function taxonomy_schema() {
  $schema['taxonomy_term_data'] = array(
    'description' => 'Stores term information.',
    'fields' => array(
      'tid' => array(
        'type' => 'serial',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'description' => 'Primary Key: Unique term ID.',
      ),
      'vid' => array(
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'default' => 0,
        'description' => 'The {taxonomy_vocabulary}.vid of the vocabulary to which the term is assigned.',
      ),
      'name' => array(
        'type' => 'varchar',
        'length' => 255,
        'not null' => TRUE,
        'default' => '',
        'description' => 'The term name.',
        'translatable' => TRUE,
      ),
      'description' => array(
        'type' => 'text',
        'not null' => FALSE,
        'size' => 'big',
        'description' => 'A description of the term.',
        'translatable' => TRUE,
      ),
      'format' => array(
        'type' => 'varchar',
        'length' => 255,
        'not null' => FALSE,
        'description' => 'The {filter_format}.format of the description.',
      ),
      'weight' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 0,
        'description' => 'The weight of this term in relation to other terms.',
      ),
    ),
    'primary key' => array('tid'),
    'foreign keys' => array(
      'vocabulary' => array(
        'table' => 'taxonomy_vocabulary',
        'columns' => array('vid' => 'vid'),
      ),
    ),
    'indexes' => array(
      'taxonomy_tree' => array('vid', 'weight', 'name'),
      'vid_name' => array('vid', 'name'),
      'name' => array('name'),
    ),
  );

  $schema['taxonomy_term_hierarchy'] = array(
    'description' => 'Stores the hierarchical relationship between terms.',
    'fields' => array(
      'tid' => array(
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'default' => 0,
        'description' => 'Primary Key: The {taxonomy_term_data}.tid of the term.',
      ),
      'parent' => array(
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'default' => 0,
        'description' => "Primary Key: The {taxonomy_term_data}.tid of the term's parent. 0 indicates no parent.",
      ),
    ),
    'indexes' => array(
      'parent' => array('parent'),
    ),
    'foreign keys' => array(
      'taxonomy_term_data' => array(
        'table' => 'taxonomy_term_data',
        'columns' => array('tid' => 'tid'),
      ),
    ),
    'primary key' => array('tid', 'parent'),
  );

  $schema['taxonomy_vocabulary'] = array(
    'description' => 'Stores vocabulary information.',
    'fields' => array(
      'vid' => array(
        'type' => 'serial',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'description' => 'Primary Key: Unique vocabulary ID.',
      ),
      'name' => array(
        'type' => 'varchar',
        'length' => 255,
        'not null' => TRUE,
        'default' => '',
        'description' => 'Name of the vocabulary.',
        'translatable' => TRUE,
      ),
      'machine_name' => array(
        'type' => 'varchar',
        'length' => 255,
        'not null' => TRUE,
        'default' => '',
        'description' => 'The vocabulary machine name.',
      ),
      'description' => array(
        'type' => 'text',
        'not null' => FALSE,
        'size' => 'big',
        'description' => 'Description of the vocabulary.',
        'translatable' => TRUE,
      ),
      'hierarchy' => array(
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'default' => 0,
        'size' => 'tiny',
        'description' => 'The type of hierarchy allowed within the vocabulary. (0 = disabled, 1 = single, 2 = multiple)',
      ),
      'module' => array(
        'type' => 'varchar',
        'length' => 255,
        'not null' => TRUE,
        'default' => '',
        'description' => 'The module which created the vocabulary.',
      ),
      'weight' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 0,
        'description' => 'The weight of this vocabulary in relation to other vocabularies.',
      ),
    ),
    'primary key' => array('vid'),
    'indexes' => array(
      'list' => array('weight', 'name'),
    ),
    'unique keys' => array(
      'machine_name' => array('machine_name'),
    ),
  );

  $schema['taxonomy_index'] = array(
    'description' => 'Maintains denormalized information about node/term relationships.',
    'fields' => array(
      'nid' => array(
        'description' => 'The {node}.nid this record tracks.',
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'default' => 0,
      ),
      'tid' => array(
         'description' => 'The term ID.',
         'type' => 'int',
         'unsigned' => TRUE,
         'not null' => TRUE,
         'default' => 0,
      ),
      'sticky' => array(
        'description' => 'Boolean indicating whether the node is sticky.',
        'type' => 'int',
        'not null' => FALSE,
        'default' => 0,
        'size' => 'tiny',
      ),
      'created' => array(
        'description' => 'The Unix timestamp when the node was created.',
        'type' => 'int',
        'not null' => TRUE,
        'default'=> 0,
      ),
    ),
    'indexes' => array(
      'term_node' => array('tid', 'sticky', 'created'),
      'nid' => array('nid'),
    ),
    'foreign keys' => array(
      'tracked_node' => array(
        'table' => 'node',
        'columns' => array('nid' => 'nid'),
      ),
      'term' => array(
        'table' => 'taxonomy_term_data',
        'columns' => array('tid' => 'tid'),
      ),
    ),
  );

  return $schema;
}

/**
 * Implements hook_field_schema().
 */
function taxonomy_field_schema($field) {
  return array(
    'columns' => array(
      'tid' => array(
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => FALSE,
      ),
    ),
    'indexes' => array(
      'tid' => array('tid'),
    ),
    'foreign keys' => array(
      'tid' => array(
        'table' => 'taxonomy_term_data',
        'columns' => array('tid' => 'tid'),
      ),
    ),
  );
}

/**
 * Implements hook_update_dependencies().
 */
function taxonomy_update_dependencies() {
  // taxonomy_update_7004() migrates taxonomy term data to fields and therefore
  // must run after all Field modules have been enabled, which happens in
  // system_update_7027().
  $dependencies['taxonomy'][7004] = array(
    'system' => 7027,
  );

  return $dependencies;
}

/**
 * Utility function: get the list of vocabularies directly from the database.
 *
 * This function is valid for a database schema version 7002.
 *
 * @ingroup update_api
 */
function _update_7002_taxonomy_get_vocabularies() {
  return db_query('SELECT v.* FROM {taxonomy_vocabulary} v ORDER BY v.weight, v.name')->fetchAllAssoc('vid', PDO::FETCH_OBJ);
}

/**
 * Rename taxonomy tables.
 */
function taxonomy_update_7001() {
  db_rename_table('term_data', 'taxonomy_term_data');
  db_rename_table('term_hierarchy', 'taxonomy_term_hierarchy');
  db_rename_table('term_node', 'taxonomy_term_node');
  db_rename_table('term_relation', 'taxonomy_term_relation');
  db_rename_table('term_synonym', 'taxonomy_term_synonym');
  db_rename_table('vocabulary', 'taxonomy_vocabulary');
  db_rename_table('vocabulary_node_types', 'taxonomy_vocabulary_node_type');
}

/**
 * Add {vocabulary}.machine_name column.
 */
function taxonomy_update_7002() {
  $field = array(
    'type' => 'varchar',
    'length' => 255,
    'not null' => TRUE,
    'default' => '',
    'description' => 'The vocabulary machine name.',
  );

  db_add_field('taxonomy_vocabulary', 'machine_name', $field);

  // Do a direct query here, rather than calling taxonomy_get_vocabularies(),
  // in case Taxonomy module is disabled.
  $vids = db_query('SELECT vid FROM {taxonomy_vocabulary}')->fetchCol();
  foreach ($vids as $vid) {
    $machine_name = 'vocabulary_' . $vid;
    db_update('taxonomy_vocabulary')
      ->fields(array('machine_name' => $machine_name))
      ->condition('vid', $vid)
      ->execute();
  }

  // The machine_name unique key can only be added after we ensure the
  // machine_name column contains unique values.
  db_add_unique_key('taxonomy_vocabulary', 'machine_name', array('machine_name'));
}

/**
 * Remove the related terms setting from vocabularies.
 *
 * This setting has not been used since Drupal 6. The {taxonomy_relations} table
 * itself is retained to allow for data to be upgraded.
 */
function taxonomy_update_7003() {
  db_drop_field('taxonomy_vocabulary', 'relations');
}

/**
 * Move taxonomy vocabulary associations for nodes to fields and field instances.
 */
function taxonomy_update_7004() {
  $taxonomy_index = array(
    'description' => 'Maintains denormalized information about node/term relationships.',
    'fields' => array(
      'nid' => array(
        'description' => 'The {node}.nid this record tracks.',
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'default' => 0,
      ),
      'tid' => array(
         'description' => 'The term ID.',
         'type' => 'int',
         'unsigned' => TRUE,
         'not null' => TRUE,
         'default' => 0,
      ),
      'sticky' => array(
        'description' => 'Boolean indicating whether the node is sticky.',
        'type' => 'int',
        'not null' => FALSE,
        'default' => 0,
        'size' => 'tiny',
      ),
      'created' => array(
        'description' => 'The Unix timestamp when the node was created.',
        'type' => 'int',
        'unsigned' => TRUE,
        'not null' => TRUE,
        'default'=> 0,
      ),
    ),
    'indexes' => array(
      'term_node' => array('tid', 'sticky', 'created'),
      'nid' => array('nid'),
    ),
    'foreign keys' => array(
      'tracked_node' => array(
        'table' => 'node',
        'columns' => array('nid' => 'nid'),
      ),
      'term' => array(
        'table' => 'taxonomy_term_data',
        'columns' => array('tid' => 'tid'),
      ),
    ),
  );
  db_create_table('taxonomy_index', $taxonomy_index);

  // Use an inline version of Drupal 6 taxonomy_get_vocabularies() here since
  // we can no longer rely on $vocabulary->nodes from the API function.
  $result = db_query('SELECT v.*, n.type FROM {taxonomy_vocabulary} v LEFT JOIN {taxonomy_vocabulary_node_type} n ON v.vid = n.vid ORDER BY v.weight, v.name');
  $vocabularies = array();
  foreach ($result as $record) {
    // If no node types are associated with a vocabulary, the LEFT JOIN will
    // return a NULL value for type.
    if (isset($record->type)) {
      $node_types[$record->vid][$record->type] = $record->type;
      unset($record->type);
      $record->nodes = $node_types[$record->vid];
    }
    elseif (!isset($record->nodes)) {
      $record->nodes = array();
    }
    $vocabularies[$record->vid] = $record;
  }

  foreach ($vocabularies as $vocabulary) {
    $field_name = 'taxonomy_' . $vocabulary->machine_name;
    $field = array(
      'field_name' => $field_name,
      'module' => 'taxonomy',
      'type' => 'taxonomy_term_reference',
      'cardinality' => $vocabulary->multiple || $vocabulary->tags ? FIELD_CARDINALITY_UNLIMITED : 1,
      'settings' => array(
        'required' => $vocabulary->required ? TRUE : FALSE,
        'allowed_values' => array(
          array(
            'vocabulary' => $vocabulary->machine_name,
            'parent' => 0,
          ),
        ),
      ),
    );
    _update_7000_field_create_field($field);

    foreach ($vocabulary->nodes as $bundle) {
      $instance = array(
        'label' => $vocabulary->name,
        'field_name' => $field_name,
        'bundle' => $bundle,
        'entity_type' => 'node',
        'settings' => array(),
        'description' => $vocabulary->help,
        'required' => $vocabulary->required,
        'widget' => array(),
        'display' => array(
          'default' => array(
            'type' => 'taxonomy_term_reference_link',
            'weight' => 10,
          ),
          'teaser' => array(
            'type' => 'taxonomy_term_reference_link',
            'weight' => 10,
          ),
        ),
      );
      if ($vocabulary->tags) {
        $instance['widget'] = array(
          'type' => 'taxonomy_autocomplete',
          'module' => 'taxonomy',
          'settings' => array(
            'size' => 60,
            'autocomplete_path' => 'taxonomy/autocomplete',
          ),
        );
      }
      else {
        $instance['widget'] = array(
          'type' => 'select',
          'module' => 'options',
          'settings' => array(),
        );
      }
      _update_7000_field_create_instance($field, $instance);
    }
  }

  // Some contrib projects stored term node associations without regard for the
  // selections in the taxonomy_vocabulary_node_types table, or have more terms
  // for a single node than the vocabulary allowed. We construct the
  // taxonomyextra field to store all the extra stuff.

  // Allowed values for this extra vocabs field is every vocabulary.
  $allowed_values = array();
  foreach (_update_7002_taxonomy_get_vocabularies() as $vocabulary) {
    $allowed_values[] = array(
      'vocabulary' => $vocabulary->machine_name,
      'parent' => 0,
    );
  }

  $field_name = 'taxonomyextra';
  $field = array(
    'field_name' => $field_name,
    'module' => 'taxonomy',
    'type' => 'taxonomy_term_reference',
    'cardinality' => FIELD_CARDINALITY_UNLIMITED,
    'settings' => array(
      'required' => FALSE,
      'allowed_values' => $allowed_values,
    ),
  );
  _update_7000_field_create_field($field);

  foreach (_update_7000_node_get_types() as $bundle) {
    $instance = array(
      'label' => 'Taxonomy upgrade extras',
      'field_name' => $field_name,
      'entity_type' => 'node',
      'bundle' => $bundle->type,
      'settings' => array(),
      'description' => 'Debris left over after upgrade from Drupal 6',
      'widget' => array(
        'type' => 'taxonomy_autocomplete',
        'module' => 'taxonomy',
        'settings' => array(),
      ),
      'display' => array(
        'default' => array(
          'type' => 'taxonomy_term_reference_link',
          'weight' => 10,
        ),
        'teaser' => array(
          'type' => 'taxonomy_term_reference_link',
          'weight' => 10,
        ),
      ),
    );
    _update_7000_field_create_instance($field, $instance);
  }

  $fields = array('help', 'multiple', 'required', 'tags');
  foreach ($fields as $field) {
    db_drop_field('taxonomy_vocabulary', $field);
  }
}

/**
 * Migrate {taxonomy_term_node} table to field storage.
 *
 * @todo: This function can possibly be made much faster by wrapping a
 * transaction around all the inserts.
 */
function taxonomy_update_7005(&$sandbox) {
  // $sandbox contents:
  // - total: The total number of term_node relationships to migrate.
  // - count: The number of term_node relationships that have been
  //   migrated so far.
  // - last: The db_query_range() offset to use when querying
  //   term_node; this field is incremented in quantities of $batch
  //   (1000) but at the end of each call to this function, last and
  //   count are the same.
  // - vocabularies: An associative array mapping vocabulary id and node
  //   type to field name. If a voc id/node type pair does not appear
  //   in this array but a term_node relationship exists mapping a
  //   term in voc id to node of that type, the relationship is
  //   assigned to the taxonomymyextra field which allows terms of all
  //   vocabularies.
  // - cursor[values], cursor[deltas]: The contents of $values and
  //   $deltas at the end of the previous call to this function. These
  //   need to be preserved across calls because a single batch of
  //   1000 rows from term_node may end in the middle of the terms for
  //   a single node revision.
  //
  // $values is the array of values about to be/most recently inserted
  // into the SQL data table for the taxonomy_term_reference
  // field. Before $values is constructed for each record, the
  // $values from the previous insert is checked to see if the two
  // records are for the same node revision id; this enables knowing
  // when to reset the delta counters which are incremented across all
  // terms for a single field on a single revision, but reset for each
  // new field and revision.
  //
  // $deltas is an associative array mapping field name to the number
  // of term references stored so far for the current revision, which
  // provides the delta value for each term reference data insert. The
  // deltas are reset for each new revision.
  
  $conditions = array(
    'type' => 'taxonomy_term_reference',
    'deleted' => 0,
  );
  $field_info = _update_7000_field_read_fields($conditions, 'field_name');

  // This is a multi-pass update. On the first call we need to initialize some
  // variables.
  if (!isset($sandbox['total'])) {
    $sandbox['last'] = 0;
    $sandbox['count'] = 0;

    // Run the same joins as the query that is used later to retrieve the
    // term_node data, this ensures that bad records in that table - for
    // tids which aren't in taxonomy_term_data or nids which aren't in {node}
    // are not included in the count.
    $sandbox['total'] = db_query('SELECT COUNT(*) FROM {taxonomy_term_data} td INNER JOIN {taxonomy_term_node} tn ON td.tid = tn.tid INNER JOIN {node} n ON tn.nid = n.nid LEFT JOIN {node} n2 ON tn.vid = n2.vid')->fetchField();

    // Use an inline version of Drupal 6 taxonomy_get_vocabularies() here since
    // we can no longer rely on $vocabulary->nodes from the API function.
    $result = db_query('SELECT v.vid, v.machine_name, n.type FROM {taxonomy_vocabulary} v INNER JOIN {taxonomy_vocabulary_node_type} n ON v.vid = n.vid');
    $vocabularies = array();
    foreach ($result as $record) {

      // If no node types are associated with a vocabulary, the LEFT JOIN will
      // return a NULL value for type.
      if (isset($record->type)) {
        $vocabularies[$record->vid][$record->type] = 'taxonomy_'. $record->machine_name;
      }
    }

    if (!empty($vocabularies)) {
      $sandbox['vocabularies'] = $vocabularies;
    }

    db_create_table('taxonomy_update_7005', array(
      'description' => 'Stores temporary data for taxonomy_update_7005.',
      'fields' => array(
        'n' => array(
          'description' => 'Preserve order.',
          'type' => 'serial',
          'unsigned' => TRUE,
          'not null' => TRUE,
        ),
        'vocab_id' => array(
          'type' => 'int',
          'unsigned' => TRUE,
          'not null' => TRUE,
          'default' => 0,
        ),
        'tid' => array(
          'type' => 'int',
          'unsigned' => TRUE,
          'not null' => TRUE,
        ),
        'nid' => array(
          'type' => 'int',
          'unsigned' => TRUE,
          'not null' => TRUE,
        ),
        'vid' => array(
          'type' => 'int',
          'unsigned' => TRUE,
          'not null' => FALSE,
          'default' => NULL,
        ),
        'type' => array(
          'type' => 'varchar',
          'length' => 32,
          'not null' => TRUE,
          'default' => '',
        ),
        'created' => array(
          'type' => 'int',
          'not null' => FALSE,
        ),
        'sticky' => array(
          'type' => 'int',
          'not null' => FALSE,
        ),
        'status' => array(
          'type' => 'int',
          'not null' => FALSE,
        ),
        'is_current' => array(
          'type' => 'int',
          'unsigned' => TRUE,
          'not null' => FALSE,
        ),
      ),
      'primary key' => array('n'),
    ));

    // Query selects all revisions at once and processes them in revision and
    // term weight order.
    $query = db_select('taxonomy_term_data', 'td');
    // We are migrating term-node relationships. If there are none for a
    // term, we do not need the term_data row.
    $query->join('taxonomy_term_node', 'tn', 'td.tid = tn.tid');
    // If a term-node relationship exists for a nid that does not exist, we
    // cannot migrate it as we have no node to relate it to; thus we do not
    // need that row from term_node.
    $query->join('node', 'n', 'tn.nid = n.nid');
    // If the current term-node relationship is for the current revision of
    // the node, this left join will match and is_current will be non-NULL
    // (we also get the current sticky and created in this case). This
    // tells us whether to insert into the current data tables in addition
    // to the revision data tables.
    $query->leftJoin('node', 'n2', 'tn.vid = n2.vid');
    $query->addField('td', 'vid', 'vocab_id');
    $query->addField('td', 'tid');
    $query->addField('tn', 'nid');
    $query->addField('tn', 'vid');
    $query->addField('n', 'type');
    $query->addField('n2', 'created');
    $query->addField('n2', 'sticky');
    $query->addField('n2', 'status');
    $query->addField('n2', 'nid', 'is_current');
    // This query must return a consistent ordering across multiple calls.
    // We need them ordered by node vid (since we use that to decide when
    // to reset the delta counters) and by term weight so they appear
    // within each node in weight order. However, tn.vid,td.weight is not
    // guaranteed to be unique, so we add tn.tid as an additional sort key
    // because tn.tid,tn.vid is the primary key of the D6 term_node table
    // and so is guaranteed unique. Unfortunately it also happens to be in
    // the wrong order which is less efficient, but c'est la vie.
    $query->orderBy('tn.vid');
    $query->orderBy('td.weight');
    $query->orderBy('tn.tid');

    // Work around a bug in the PostgreSQL driver that would result in fatal
    // errors when this subquery is used in the insert query below. See
    // https://drupal.org/node/2057693.
    $fields = &$query->getFields();
    unset($fields['td.weight']);
    unset($fields['tn.tid']);

    db_insert('taxonomy_update_7005')
      ->from($query)
      ->execute();
  }
  else {
    // We do each pass in batches of 1000.
    $batch = 1000;

    $result = db_query_range('SELECT vocab_id, tid, nid, vid, type, created, sticky, status, is_current FROM {taxonomy_update_7005} ORDER BY n', $sandbox['last'], $batch);
    if (isset($sandbox['cursor'])) {
      $values = $sandbox['cursor']['values'];
      $deltas = $sandbox['cursor']['deltas'];
    }
    else {
      $deltas = array();
    }
    foreach ($result as $record) {
      $sandbox['count'] += 1;

      // Use the valid field for this vocabulary and node type or use the
      // overflow vocabulary if there is no valid field.
      $field_name = isset($sandbox['vocabularies'][$record->vocab_id][$record->type]) ? $sandbox['vocabularies'][$record->vocab_id][$record->type] : 'taxonomyextra';
      $field = $field_info[$field_name];

      // Start deltas from 0, and increment by one for each term attached to a
      // node.
      if (!isset($deltas[$field_name])) {
        $deltas[$field_name] = 0;
      }

      if (isset($values)) {

        // If the last inserted revision_id is the same as the current record,
        // use the previous deltas to calculate the next delta.
        if ($record->vid == $values[2]) {

          // For limited cardinality fields, the delta must not be allowed to
          // exceed the cardinality during the update. So ensure that the
          // delta about to be inserted is within this limit.
          // @see field_default_validate().
          if ($field['cardinality'] != FIELD_CARDINALITY_UNLIMITED && ($deltas[$field_name] + 1) > $field['cardinality']) {

            // For excess values of a single-term vocabulary, switch over to
            // the overflow field.
            $field_name = 'taxonomyextra';
            $field = $field_info[$field_name];
            if (!isset($deltas[$field_name])) {
              $deltas[$field_name] = 0;
            }
          }
        }
        else {

          // When the record is a new revision, empty the deltas array.
          $deltas = array($field_name => 0);
        }
      }

      // Table and column found in the field's storage details. During upgrades,
      // it's always SQL.
      $table_name = "field_data_{$field_name}";
      $revision_name = "field_revision_{$field_name}";
      $value_column = $field_name . '_tid';

      // Column names and values in field storage are the same for current and
      // revision.
      $columns = array('entity_type', 'entity_id', 'revision_id', 'bundle', 'language', 'delta', $value_column);
      $values = array('node', $record->nid, $record->vid, $record->type, LANGUAGE_NONE, $deltas[$field_name]++, $record->tid);

      // Insert rows into the revision table.
      db_insert($revision_name)->fields($columns)->values($values)->execute();

      // is_current column is a node ID if this revision is also current.
      if ($record->is_current) {
        db_insert($table_name)->fields($columns)->values($values)->execute();
        // Only insert a record in the taxonomy index if the node is published.
        if ($record->status) {
          // Update the {taxonomy_index} table.
          db_insert('taxonomy_index')
            ->fields(array('nid', 'tid', 'sticky', 'created',))
            ->values(array($record->nid, $record->tid, $record->sticky, $record->created))
            ->execute();
        }
      }
    }

    // Store the set of inserted values and the current revision's deltas in the
    // sandbox.
    $sandbox['cursor'] = array(
      'values' => $values,
      'deltas' => $deltas,
    );
    $sandbox['last'] += $batch;
  }

  if ($sandbox['count'] < $sandbox['total']) {
    $sandbox['#finished'] = FALSE;
  }
  else {
    db_drop_table('taxonomy_vocabulary_node_type');
    db_drop_table('taxonomy_term_node');

    // If there are no vocabs, we're done.
    db_drop_table('taxonomy_update_7005');
    $sandbox['#finished'] = TRUE;

    // Determine necessity of taxonomyextras field.
    $field = $field_info['taxonomyextra'];
    $revision_name = 'field_revision_' . $field['field_name'];
    $node_types = db_select($revision_name)->distinct()->fields($revision_name, array('bundle'))
      ->execute()->fetchCol();

    if (empty($node_types)) {
      // Delete the overflow field if there are no rows in the revision table.
      _update_7000_field_delete_field('taxonomyextra');
    }
    else {
      // Remove instances which are not actually used.
      $bundles = db_query('SELECT bundle FROM {field_config_instance} WHERE field_name = :field_name', array(':field_name' => 'taxonomyextra'))->fetchCol();
      $bundles = array_diff($bundles, $node_types);
      foreach ($bundles as $bundle) {
        _update_7000_field_delete_instance('taxonomyextra', 'node', $bundle);
      }
    }
  }
}

/**
 * Add {taxonomy_term_data}.format column.
 */
function taxonomy_update_7006() {
  db_add_field('taxonomy_term_data', 'format', array(
    'type' => 'int',
    'unsigned' => TRUE,
    'not null' => FALSE,
    'description' => 'The {filter_format}.format of the description.',
  ));
}

/**
 * Add index on {taxonomy_term_data}.name column to speed up taxonomy_get_term_by_name().
 */
function taxonomy_update_7007() {
  db_add_index('taxonomy_term_data', 'name', array('name'));
}

/**
 * Change the weight columns to normal int.
 */
function taxonomy_update_7008() {
  db_drop_index('taxonomy_term_data', 'taxonomy_tree');
  db_change_field('taxonomy_term_data', 'weight', 'weight', array(
    'type' => 'int',
    'not null' => TRUE,
    'default' => 0,
    'description' => 'The weight of this term in relation to other terms.',
  ), array(
    'indexes' => array(
       'taxonomy_tree' => array('vid', 'weight', 'name'),
    ),
  ));

  db_drop_index('taxonomy_vocabulary', 'list');
  db_change_field('taxonomy_vocabulary', 'weight', 'weight', array(
    'type' => 'int',
    'not null' => TRUE,
    'default' => 0,
    'description' => 'The weight of this vocabulary in relation to other vocabularies.',
  ), array(
    'indexes' => array(
      'list' => array('weight', 'name'),
    ),
  ));
}

/**
 * Change {taxonomy_term_data}.format into varchar.
 */
function taxonomy_update_7009() {
  db_change_field('taxonomy_term_data', 'format', 'format', array(
    'type' => 'varchar',
    'length' => 255,
    'not null' => FALSE,
    'description' => 'The {filter_format}.format of the description.',
  ));
}

/**
 * Change {taxonomy_index}.created to support signed int.
*/
function taxonomy_update_7010() {
  db_change_field('taxonomy_index', 'created', 'created', array(
    'description' => 'The Unix timestamp when the node was created.',
    'type' => 'int',
    'unsigned' => FALSE,
    'not null' => TRUE,
    'default'=> 0,
  ));
}

/**
 * @addtogroup updates-7.x-extra
 * @{
 */

/**
 * Drop unpublished nodes from the index.
 */
function taxonomy_update_7011(&$sandbox) {
  // Initialize information needed by the batch update system.
  if (!isset($sandbox['progress'])) {
    $sandbox['progress'] = 0;
    $sandbox['max'] = db_query('SELECT COUNT(DISTINCT n.nid) FROM {node} n INNER JOIN {taxonomy_index} t ON n.nid = t.nid WHERE n.status = :status', array(':status' => NODE_NOT_PUBLISHED))->fetchField();
    // If there's no data, don't bother with the extra work.
    if (empty($sandbox['max'])) {
      return;
    }
  }

  // Process records in groups of 5000.
  $limit = 5000;
  $nids = db_query_range('SELECT DISTINCT n.nid FROM {node} n INNER JOIN {taxonomy_index} t ON n.nid = t.nid WHERE n.status = :status', 0, $limit, array(':status' => NODE_NOT_PUBLISHED))->fetchCol();
  if (!empty($nids)) {
    db_delete('taxonomy_index')
      ->condition('nid', $nids)
      ->execute();
  }

  // Update our progress information for the batch update.
  $sandbox['progress'] += $limit;

  // Indicate our current progress to the batch update system, if the update is
  // not yet complete.
  if ($sandbox['progress'] < $sandbox['max']) {
    $sandbox['#finished'] = $sandbox['progress'] / $sandbox['max'];
  }
}

/**
 * @} End of "addtogroup updates-7.x-extra".
 */