magento, url, rewrite

urls.md

How product Url is generated

On product list page to get product urls Magento performs the following query to assign request_path to every item in product collection:

SELECT
  `e`.`entity_id`                                           AS `product_id`,
  IFNULL(url_rewrite.request_path, default_ur.request_path) AS `request_path`
FROM `catalog_product_entity` AS `e`
  LEFT JOIN `enterprise_catalog_product_rewrite` AS `url_rewrite_product`
    ON url_rewrite_product.product_id = e.entity_id AND url_rewrite_product.store_id = 1
  LEFT JOIN `enterprise_url_rewrite` AS `url_rewrite`
    ON url_rewrite_product.url_rewrite_id = url_rewrite.url_rewrite_id AND url_rewrite.is_system = 1
  LEFT JOIN `enterprise_catalog_product_rewrite` AS `default_urp`
    ON default_urp.product_id = e.entity_id AND default_urp.store_id = 0
  LEFT JOIN `enterprise_url_rewrite` AS `default_ur` ON default_ur.url_rewrite_id = default_urp.url_rewrite_id
WHERE (e.entity_id IN ('549', '551', '552', '553', '554', '555'))

sample output:

product_id	request_path
549	blue-horizons-bracelets
551	pearl-stud-earrings
552	swing-time-earrings
553	silver-desert-necklace
554	swiss-movement-sports-watch-h
555	pearl-necklace-set-test

Then Magento checks if there is current_category in registry to get category_id, the same time it checks if category path should be used in Product URLs at all. If it should be used then category 'request_path' is taken from DB as result of using \Enterprise_Catalog_Model_Resource_Category::loadByCategory method, query looks like this:

SELECT
  IFNULL(url_rewrite_cat.id, default_urc.id)                AS `id`,
  IFNULL(url_rewrite.request_path, default_ur.request_path) AS `request_path`
FROM `catalog_category_entity` AS `main_table`
  LEFT JOIN `enterprise_catalog_category_rewrite` AS `url_rewrite_cat`
    ON url_rewrite_cat.category_id = main_table.entity_id AND url_rewrite_cat.store_id = 1
  LEFT JOIN `enterprise_url_rewrite` AS `url_rewrite` ON url_rewrite.url_rewrite_id = url_rewrite_cat.url_rewrite_id
  LEFT JOIN `enterprise_catalog_category_rewrite` AS `default_urc`
    ON default_urc.category_id = main_table.entity_id AND default_urc.store_id = 0
  LEFT JOIN `enterprise_url_rewrite` AS `default_ur` ON default_ur.url_rewrite_id = default_urc.url_rewrite_id
WHERE (main_table.entity_id = 19)

sample output:

id	request path
34	accessories/jewelry

Request path is being concatenated with product's 'request_path'

 $requestPath = $categoryRewrite->getRequestPath() . '/' . $requestPath;
 ...
 $product->setRequestPath($requestPath);

The result string is product's URL

Reversed transformation

In \Mage_Core_Controller_Varien_Front::dispatch method there is a line of code that rewrite process start with:

$this->_getRequestRewriteController()->rewrite();

Request path is divided on pathes using '/'. Rewrites are being loaded for those pathes from enterprise_url_rewrite table. Sample query:

SELECT `m`.*
FROM `enterprise_url_rewrite` AS `m`
WHERE (m.request_path IN
       ('accessories/jewelry/swiss-movement-sports-watch-h.html', 'swiss-movement-sports-watch-h', 'accessories/jewelry/swiss-movement-sports-watch-h'))
ORDER BY `m`.`store_id` DESC

Result rows then are being sorted by priority using special matcher class. If rewrite is not system then Magento tries to get redirect for this rewrite, if there is a RP option for redirect than appropriate headers will be sent for redirection. Rewrite object has target_path field that will be used to determine controller, action and product id.