Question

我有一张包含以下数据的表

(1995, 'a'),
(1996, 'a'),
(1997, 'a'),
(1998, 'a'),
(2000, 'a'),

(1995, 'b'),
(1997, 'b'),
(1998, 'b'),

(1995, 'c'),
(1999, 'c'),
(2000, 'c'),

(1999, 'd'),

(1999, 'e'),

(1999, 'f');

我需要计算每个字母表的最大序列长度，例如：因为'a'在1995,1996,1997,1998中连续存在，'a'的最大序列长度为4.预期输出为：

a-4
b-2
c-2
d-1
e-1
f-1

不确定如何继续。

由于

Answer 1

请考虑以下事项：

DROP TABLE IF EXISTS my_table;

CREATE TABLE my_table (year INT NOT NULL,string CHAR(1) NOT NULL,PRIMARY KEY(year,string));

INSERT INTO my_table VALUES
    (1995, 'a'),
    (1996, 'a'),
    (1997, 'a'),
    (1998, 'a'),
    (2000, 'a'),

    (1995, 'b'),
    (1997, 'b'),
    (1998, 'b'),

    (1995, 'c'),
    (1999, 'c'),
    (2000, 'c'),

    (1999, 'd'),

    (1999, 'e'),

    (1999, 'f');

SELECT string
     , MAX(diff) FROM
     ( SELECT a.string
            , a.year start
            , MIN(c.year) end 
            , MIN(c.year) - a.year+1 diff
         FROM my_table a
         LEFT 
         JOIN my_table b
           ON b.string = a.string
          AND b.year + 1 = a.year
         LEFT 
         JOIN my_table c
           ON c.string = a.string
          AND c.year >= a.year
         LEFT 
         JOIN my_table d
           ON d.string = a.string
          AND d.year - 1 = c.year
        WHERE b.string IS NULL 
          AND c.string IS NOT NULL
          AND d.string IS NULL
        GROUP 
           BY a.string,a.year
     ) a 
 GROUP 
    BY string; 

+--------+-----------+
| string | MAX(diff) |
+--------+-----------+
| a      |         4 |
| b      |         2 |
| c      |         2 |
| d      |         1 |
| e      |         1 |
| f      |         1 |
+--------+-----------+

编辑：我对变量有一种审美厌恶，但事实上，在表现方面，他们的效率会更高......

SELECT string
     , MAX(i) 
  FROM
     (
       SELECT year
            , string
            , IF(year=@prev_year+1,IF(string=@prev_string,@i:=@i+1,@i:=1),@i:=1)i
            , @prev_year := year
            , @prev_string := string 
         FROM my_table
            , ( SELECT @prev_year=0,@prev_string:='',@i:=1 ) vars 
        ORDER 
           BY string
            , year
     ) x
 GROUP 
    BY string;
+--------+--------+
| string | MAX(i) |
+--------+--------+
| a      |      4 |
| b      |      2 |
| c      |      2 |
| d      |      1 |
| e      |      1 |
| f      |      1 |
+--------+--------+

Answer 2

我认为下面的代码可以帮助你。

SELECT string, COUNT(string) AS MAX
FROM my_table
GROUP BY string
HAVING COUNT(string)>=1;

Mysql：计算最大序列长度

2 个答案: