从表中检索所有数据，然后计算另一个中的结果数

Question

I am wondering if this is possible using just one MySQL query.

What I am making is a tagging system. Therefore, I have two tables:

Tags, id | heading

Tagged, id | image_id | tag_id | timestamp | strength

Everything works fine, but now I am getting to the stage where I am displaying the tags to my users and plan to sort them by how many times a tag has been used.

So in a nutshell, I need something that will do the following,

1. Grab all tags for the current image displayed, eg id = 34 .
2. Count how many times the tag has been used.
3. Order the tag list by these results, going from high to low.

Now I have tried looking into joins, left joins, etc., but it all looks very overwhelming and I am finding it hard to get a grasp of it.

If anyone could explain how to do this easily, that would be great!

EDIT

I currently have this

SELECT tags.id, COUNT( tagged.id ) AS count
FROM tags
INNER JOIN tagged ON tags.id = tagged.tag_id
WHERE tagged.image_id =  '1'
GROUP BY tags.id
ORDER BY count DESC 

This only returns how many times a tag has been tagged to the current image_id, how would i make it so that it counts from all images?

Answer 1

You can use GROUP BY to group together the tags that are the same, and use COUNT() to count them:

  SELECT tags.id  as id,
         COUNT(1) as cnt
    FROM tags INNER JOIN tagged
      ON tags.id = tagged.tag_id
GROUP BY tags.id
ORDER BY cnt DESC

Here's what it does exactly:

• The join takes each entry in tagged, and pairs it with its corresponding tag.
• We then group by the tag id, so that we only get one entry for each id.
• The COUNT function counts how many elements are in each group.
• We finally sort it in descending order by the count, giving the most popular tags on top.

In addition, if you want each image to be able to be tagged with multiple tags, you'll want to move the image information to its own table, and create a table solely linking the two together.

Answer 2

You can optimize your SQL query to do the work for you:

SELECT tags.id, COUNT(id) AS count
FROM tags
INNER JOIN tagged ON tags.id = tagged.tag_id
WHERE tagged.image_id = 34;
GROUP BY tags.id

This will select every tag which is set to be tagged on image id 34 and then fetch the count of total number of images using that tag. Visual explanation of JOINs

Answer 3

You are probably at the point where you should have an intermediate table which holds the relationships between the two tables, usually termed " normalization ".

tag_image_table
=========

tag | image
===========
1 | 23
99| 23

which represents image #23 associated to 2 tags.

This makes it much easier to select all images associated to a tag, and all tags associated to an image.

Answer 4

To get the global usage of the tags, you probably need to rejoin to the tagged table, since you are applying the where condition on the original tagged table. Try this:

SELECT 
    t2.id,
    COUNT(t3.tag_id) AS global_count,
    GROUP_CONCAT(t3.image_id) AS images_with_tag
FROM 
    tagged t1
    JOIN tags t2 ON t2.id = t1.tag_id
    LEFT JOIN tagged t3 ON t3.tag_id = t2.id
WHERE t1.image_id = 1
GROUP BY t2.id
ORDER BY global_count DESC