function hook_ranking
Same name in other branches
- 7.x modules/node/node.api.php \hook_ranking()
- 8.9.x core/modules/node/node.api.php \hook_ranking()
- 10 core/modules/node/node.api.php \hook_ranking()
- 11.x core/modules/node/node.api.php \hook_ranking()
Provide additional methods of scoring for core search results for nodes.
A node's search score is used to rank it among other nodes matched by the search, with the highest-ranked nodes appearing first in the search listing.
For example, a module allowing users to vote on content could expose an option to allow search results' rankings to be influenced by the average voting score of a node.
All scoring mechanisms are provided as options to site administrators, and may be tweaked based on individual sites or disabled altogether if they do not make sense. Individual scoring mechanisms, if enabled, are assigned a weight from 1 to 10. The weight represents the factor of magnification of the ranking mechanism, with higher-weighted ranking mechanisms having more influence. In order for the weight system to work, each scoring mechanism must return a value between 0 and 1 for every node. That value is then multiplied by the administrator-assigned weight for the ranking mechanism, and then the weighted scores from all ranking mechanisms are added, which brings about the same result as a weighted average.
Return value
array An associative array of ranking data. The keys should be strings, corresponding to the internal name of the ranking mechanism, such as 'recent', or 'comments'. The values should be arrays themselves, with the following keys available:
- title: (required) The human readable name of the ranking mechanism.
- join: (optional) An array with information to join any additional necessary table. This is not necessary if the table required is already joined to by the base query, such as for the {node} table. Other tables should use the full table name as an alias to avoid naming collisions.
- score: (required) The part of a query string to calculate the score for the ranking mechanism based on values in the database. This does not need to be wrapped in parentheses, as it will be done automatically; it also does not need to take the weighted system into account, as it will be done automatically. It does, however, need to calculate a decimal between 0 and 1; be careful not to cast the entire score to an integer by inadvertently introducing a variable argument.
- arguments: (optional) If any arguments are required for the score, they can be specified in an array here.
Related topics
3 functions implement hook_ranking()
Note: this list is generated by pattern matching, so it may include some functions that are not actually implementations of this hook.
- comment_ranking in core/
modules/ comment/ comment.module - Implements hook_ranking().
- node_ranking in core/
modules/ node/ node.module - Implements hook_ranking().
- statistics_ranking in core/
modules/ statistics/ statistics.module - Implements hook_ranking().
1 invocation of hook_ranking()
- NodeSearch::getRankings in core/
modules/ node/ src/ Plugin/ Search/ NodeSearch.php - Gathers ranking definitions from hook_ranking().
File
-
core/
modules/ node/ node.api.php, line 374
Code
function hook_ranking() {
// If voting is disabled, we can avoid returning the array, no hard feelings.
if (\Drupal::config('vote.settings')->get('node_enabled')) {
return [
'vote_average' => [
'title' => t('Average vote'),
// Note that we use i.sid, the search index's search item id, rather than
// n.nid.
'join' => [
'type' => 'LEFT',
'table' => 'vote_node_data',
'alias' => 'vote_node_data',
'on' => 'vote_node_data.nid = i.sid',
],
// The highest possible score should be 1, and the lowest possible score,
// always 0, should be 0.
'score' => 'vote_node_data.average / CAST(%f AS DECIMAL)',
// Pass in the highest possible voting score as a decimal argument.
'arguments' => [
\Drupal::config('vote.settings')->get('score_max'),
],
],
];
}
}
Buggy or inaccurate documentation? Please file an issue. Need support? Need help programming? Connect with the Drupal community.