如何使用 MySQL 识别序列中具有重复值的实体？

Question

我有一个 table:

 UNIT_ID | YEAR | MONTH |  VAR
---------+------+-------+------
    1    | 2015 |   1   |   0
    1    | 2015 |   2   |   0
    1    | 2015 |   3   |   0
    2    | 2015 |   1   |  10
    2    | 2015 |   2   |  10
    2    | 2015 |   3   |  10
    1    | 2015 |   4   |   5
    1    | 2015 |   5   |   5
    1    | 2015 |   6   |   5
    2    | 2015 |   4   |  10
    2    | 2015 |   5   |   3
    2    | 2015 |   6   |   3
    3    | 2016 |   1   |   3
    3    | 2016 |   2   |   3
    3    | 2016 |   3   |   3
    3    | 2016 |   4   |   3
    2    | 2016 |   6   |   0
    2    | 2016 |   7   |   0
    2    | 2016 |   8   |   0

我想知道哪些单位的序列大于 3 个零或大于 4 个重复值。按年份分组。所以，我的结果 table 会是这样的：

1 | 2015 | true
2 | 2015 | true
2 | 2016 | true

我找到了 this solution 但不幸的是我无法适应我的情况。我还需要查询在 MySQL.

Answer 1

你可以加入他们 4 次。最后一个连接是 left join 以允许 3 0 的大小写。

select a.unit_id, a.year, 'true'
from tbl a
join tbl b on a.unit_id = b.unit_id and a.year = b.year and a.month+1 = b.month and a.var = b.var
join tbl c on b.unit_id = c.unit_id and b.year = c.year and b.month+1 = c.month and b.var = c.var
left join tbl d on c.unit_id = d.unit_id and c.year = d.year and c.month+1 = d.month and c.var = d.var
where a.var = 0 or d.var is not null;

更快更通用的解决方案。它只扫描 table 一次，并使用用户定义的变量（@pu 表示前一个 unit_id，@py 表示前一年，等等）来记住前一行：

select distinct unit_id, year
from (
  select unit_id, `year`, `month`, `var`, 
    if(unit_id=@pu and `year`=@py and `month`=@pm+1 and `var`=@pv, @i:=@i+1, @i:=1)*
    if(@pu:=unit_id,1,1)*if(@py:=`year`,1,1)*if(@pm:=`month`,1,1)*if(@pv:=`var`,1,1) as c
  from table1 a
  join (select @pu:=null, @py:=null, @pm:=null, @pv:=null, @i:=1) b
  order by unit_id, `year`, `month`, `var`) a
group by unit_id, `year`, `var`
having (`var` = 0 and max(c) >= 3) or (`var` != 0 and max(c) >= 4);

fiddle

如何使用 MySQL 识别序列中具有重复值的实体？

How to identify entities which have repeated values in sequence using MySQL?

mysql

sql

sequence