在连接另一个表后对MySQL结果进行排名，并按条件过滤它们

Question

I found how to rank results, but have yet to find a solution for how to rank them after they've been filtered by a condition. What I need to do is:

1) Sort scores in descending order
2) Join the results with the account table
3) Filter the results by age and gender (from the account table)
4) Rank them
5) Grab the results within a range of ranks (IE: 25-75, but I put 0-10 for simplicity below)

Each solution I've found either doesn't properly rank them, or ranks them and then filters the results. However, this leaves me with ranks like 1, 3, 4, 6, 10, etc when I want 1, 2, 3, 4, 5, etc. or 25, 26, 27, 28, 29, etc.

Here is what I have right now:

SELECT a.id, a.username, a.country, score, user_id, duration, rank
FROM (
    SELECT s.*, a.age, a.gender, @rownum := @rownum + 1 AS rank
    FROM scores s,
        (SELECT @rownum := 0) r
    LEFT JOIN accounts a ON 1
    WHERE a.age>=18 && a.age<=35 && a.gender='m'
    ORDER BY score DESC, duration DESC, time ) `selection`
LEFT JOIN accounts a ON a.id=user_id
WHERE rank >= 0 && rank <= 10 && a.age>=18 && a.age<=35 && a.gender='m'

The problem with the above query is that it gives me correct rankings without any gaps, but the scores are completely out of order. IE. 5, 1, 10, 7, 125, 50, etc. IE. They come out in the same order they are in the database, unsorted.

This is what my previous query looked like:

SELECT a.id, a.username, a.country, score, duration, rank
FROM (
    SELECT s.*, @rownum := @rownum + 1 AS rank
    FROM scores s,
        (SELECT @rownum := 0) r
    ORDER BY score DESC, duration DESC, time ) `selection`
LEFT JOIN accounts a ON a.id=user_id
WHERE rank >= 0 && rank <= 10 && a.age>=18 && a.age<=35 && a.gender='m'

This query gives me properly sorted scores, but with ranks like 3, 5, 8, 9, 11 instead of 1, 2, 3, 4, 5.

Here is some real-world sample data...
For the sake of time, id, username, country, are all omitted because they are irrelevant data and just repeat all the way down the table. The age all the way down is '34' just as the condition shows, the country and username are the same all the way down since this particular user is the only one in the table with the age of 34.

The first query from above with proper rankings, but incorrectly sorted scores (IE. the same order as they were added to the database):

  age | score | rank
--------------------
  34  |   5   |  1
  34  |   1   |  2
  34  |  22   |  3
  34  |  13   |  4
  34  |  23   |  5
  34  |  23   |  6
  34  |  34   |  7
  34  |  32   |  8
  34  |  58   |  9
  34  |  76   |  10

The second query from above, properly sorted, but ranks still take other users into consideration when ranking, then get excluded in the final result:

  age | score | rank
---------------------
  34  |  76   |  3  --- This should have started at 1
  34  |  62   |  4
  34  |  58   |  5
  34  |  42   |  7  --- Notice 6 was skipped
  34  |  34   |  8
  34  |  32   |  9
  34  |  29   |  10

What I expect:

  age | score | rank
--------------------
  34  |  76   |  1
  34  |  62   |  2
  34  |  58   |  3
  34  |  42   |  4
  34  |  34   |  5
  34  |  32   |  6
  34  |  29   |  7

Answer 1

I can offer the following query based on your sample data set:

SET @rank = 0;
SET @score = NULL;

SELECT
    age,
    @rank:=CASE WHEN @score = score THEN @rank ELSE @rank + 1 END AS rank,
    @score:=score AS score
FROM scores
ORDER BY rank;

The computes the rank of the scores, in terms of the definition for which "rank" is usually used. If two scores be the same, then they have the same rank. If you really want the row number, then the above query can be easily modified.

Output:

Demo here:

Rextester

Answer 2

SELECT a.id, a.username, a.country, score, user_id, duration, rank
FROM (
    SELECT s.*, a.age, a.gender, @rownum := @rownum + 1 AS rank
    FROM scores s,
        (SELECT @rownum := 0) r
    LEFT JOIN accounts a ON 1
    WHERE a.age>=18 && a.age<=35 && a.gender='m'
    ORDER BY score DESC, duration DESC, time ) `selection`
LEFT JOIN accounts a ON a.id=user_id
WHERE rank >= 0 && rank <= 10 && a.age>=18