How to simplify embedded mysql query into a JOIN? - mysql

Question

I recently read something about performance problems with embedded mysql queries, so I wanted to know how I could change the following to a "JOIN" (supposedly with better performance?).
I have two tables:
CREATE TABLE IF NOT EXISTS `blog_categories` (
`category_id` int(11) NOT NULL AUTO_INCREMENT,
`category_name` varchar(300) COLLATE utf8_unicode_ci NOT NULL,
`category_name_url` varchar(300) COLLATE utf8_unicode_ci NOT NULL,
`category_status` enum('online','offline') COLLATE utf8_unicode_ci NOT NULL DEFAULT 'offline'
) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci AUTO_INCREMENT=8 ;
CREATE TABLE IF NOT EXISTS `blog_articles` (
`article_id` int(11) NOT NULL AUTO_INCREMENT,
`article_title` tinytext COLLATE utf8_unicode_ci NOT NULL,
`category_name` varchar(100) COLLATE utf8_unicode_ci NOT NULL
) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci AUTO_INCREMENT=26 ;
The logic is basically to select all the categories that have articles associated with them. Each row in the blog_articles table includes a category_name
And here is the query I'm using (which checks out and works fine):
$sql = "SELECT category_name
, category_name_url
FROM blog_categories
WHERE (
(SELECT COUNT(*)
FROM blog_articles, blog_categories
WHERE blog_articles.category_name = blog_categories.category_name
) > 0
AND blog_categories.category_status = 'online')";
I'm still new to "JOIN", and am not sure how to change it when using "COUNT(*)" in the mix too.

Answer 1

Rather than COUNT(*) > 0, use EXISTS
"SELECT category_name, category_name_url FROM blog_categories WHERE EXISTS (" .
"(SELECT 1 FROM blog_articles INNER JOIN blog_categories
ON blog_articles.category_name = blog_categories.category_name) AND " .
"blog_categories.category_status = 'online'" .
")";
This is the old style join syntax:
SELECT 1 FROM blog_articles, blog_categories
WHERE blog_articles.category_name = blog_categories.category_name
This is the ANSI form of the same:
SELECT 1
FROM blog_articles
INNER JOIN blog_categories
ON blog_articles.category_name = blog_categories.category_name
Update (in response to posters comments): This query does what you require:
"SELECT category_name, category_name_url
FROM blog_categories
WHERE category_name IN (SELECT DISTINCT category_name FROM blog_articles)
AND blog_categories.category_status = 'online'"

Answer 2

I'd use EXISTS:
$sql = "SELECT category_name
, category_name_url
FROM blog_categories
WHERE EXISTS
( SELECT *
FROM blog_articles
WHERE blog_articles.category_name = blog_categories.category_name
)
AND blog_categories.category_status = 'online'
";
or JOIN:
$sql = "SELECT c.category_name
, c.category_name_url
FROM blog_categories AS c
JOIN blog_articles AS a
ON a.category_name = c.category_name
WHERE c.category_status = 'online'
GROUP BY c.category_name
";
#Jay: I doubt your query as it is, shows only categories that have articles posted. It counts all articles from all categories and then either shows all categories (if the count is > 0) or no categories at all (if the count = 0).
The reason is that the blog_categories in your subquery is not related to the blog_categories in the main query by any condition.

Answer 3

You can try this also:
SELECT category_name, category_name_url
FROM blog_categories
WHERE category_id IN (
SELECT blog_categories.category_id
FROM blog_articles
INNER JOIN blog_categories
ON blog_categories.category_name = blog_articles.category_name
AND blog_categories.category_status = 'online'
)