MySQL 计算性别实例答案

【问题标题】：MySQL counting gender instancesMySQL 计算性别实例
【发布时间】：2013-03-16 04:36:46
【问题描述】：

我有以下 MySQL 查询。

USE demo;
SELECT 
  loc.name,
  sum(case cr_sex when 'F' then 1 else 0 end) as 'F#', 
  sum(case cr_sex when 'M' then 1 else 0 end) as 'M#'
FROM case_report_main as cr 
JOIN location as loc on cr.cr_name = loc.name
GROUP BY loc.name

这按我想要的方式工作，它计算每个性别的每次出现并按位置名称对其进行分组。但是当我尝试将它与另一个表加入时

USE demo;
SELECT 
  loc.name,
  sum(case cr_sex when 'F' then 1 else 0 end) as 'F#', 
  sum(case cr_sex when 'M' then 1 else 0 end) as 'M#'
FROM case_report_main as cr 
JOIN location as loc on cr.cr_name=loc.name
JOIN map_polygons as mp on mp.polygon_name=loc.name
GROUP BY loc.name

值完全混乱，返回数千而不是预期的数十和数百。

我认为我理解这个问题，附加表增加了行数。我不确定如何解决这个问题，或者这是否真的是问题所在。

【问题讨论】：

标签： mysql sql select join

【解决方案1】：

因为可能有记录在表map_polygons 上有多个匹配项导致对性别的计数无效，请尝试将第一个查询放在子查询中，

SELECT  mp.*,
        loc.name
        loc.`F#`
        loc.`M#`
FROM
        (
            SELECT  loc.name,
                    sum(case cr_sex when 'F' then 1 else 0 end) as 'F#', 
                    sum(case cr_sex when 'M' then 1 else 0 end) as 'M#'
            FROM case_report_main as cr 
            JOIN location as loc on cr.cr_name = loc.name
            GROUP BY loc.name
        ) loc
        INNER JOIN map_polygons mp on mp.polygon_name=loc.name

顺便说一下，既然是mysql，可以把boolean结果加起来，让它更短

SELECT  mp.*,
        loc.name
        loc.`F#`
        loc.`M#`
FROM
        (
            SELECT  loc.name,
                    sum(cr_sex = 'F') as 'F#', 
                    sum(cr_sex = 'M') as 'M#'
            FROM case_report_main as cr 
            JOIN location as loc on cr.cr_name = loc.name
            GROUP BY loc.name
        ) loc
        INNER JOIN map_polygons mp on mp.polygon_name=loc.name

【讨论】：

是的，这似乎奏效了。谢谢！我想我需要努力理解子查询。萨拉马特卡巴扬。
没问题卡巴扬。如果你需要的话，这里还有一篇关于加入的好文章。 Visual Representation of SQL Joins

【解决方案2】：

在 map_polygon 表中，你的 polygon_name 是唯一的吗？这不是唯一的，您可能需要在选择字段中进行更改

SELECT 
  loc.name,
  count(distinct case cr_sex when 'F' then cr.id else null end) as 'F#', 
  count(distinct case cr_sex when 'M' then cr.id else null end) as 'M#'
FROM case_report_main as cr 
JOIN location as loc on cr.cr_name=loc.name
JOIN map_polygons as mp on mp.polygon_name=loc.name
GROUP BY loc.name

【讨论】：