表:TeamPoints
Column Name | Type |
---|---|
team_id | int |
name | varchar |
points | int |
表:PointsChange
Column Name | Type |
---|---|
team_id | int |
points_change | int |
分数的变化分以下情况:
TeamPoints 表中出现的每一个 team_id 均会在这张表中出现。
国家队的全球排名是按 降序排列 所有队伍的得分后所得出的排名。
如果两支队伍得分相同,我们将按其名称的 字典顺序 排列以打破平衡。
每支国家队的分数应根据其相应的 points_change 进行更新。
编写解决方案来计算在分数更新后,每个队伍的全球排名的变化。
以 任意顺序 返回结果。
查询结果的格式如下例所示:
示例 1:
输入:
TeamPoints 表:
team_id | name | points |
---|---|---|
3 | Algeria | 1431 |
1 | Senegal | 2132 |
2 | New Zealand | 1402 |
4 | Croatia | 1817 |
PointsChange 表:
team_id | points_change |
---|---|
3 | 399 |
2 | 0 |
4 | 13 |
1 | -22 |
输出:
team_id | name | rank_diff |
---|---|---|
1 | Senegal | 0 |
4 | Croatia | -1 |
3 | Algeria | 1 |
2 | New Zealand | 0 |
解释:
世界排名如下所示:
team_id | name | points | rank |
---|---|---|---|
1 | Senegal | 2132 | 1 |
4 | Croatia | 1817 | 2 |
3 | Algeria | 1431 | 3 |
2 | New Zealand | 1402 | 4 |
在更新分数后,世界排名变为下表:
team_id | name | points | rank |
---|---|---|---|
1 | Senegal | 2110 | 1 |
3 | Algeria | 1830 | 2 |
4 | Croatia | 1830 | 3 |
2 | New Zealand | 1402 | 4 |
由于在更新分数后,Algeria 和 Croatia 的得分相同,因此根据字典顺序对它们进行排序。
Senegal 丢失了 22分 但他们的排名没有改变。
Croatia 获得了 13分 但是他们的排名下降了 1 名。
Algeria 获得 399分,排名上升了 1 名。
New Zealand 没有获得或丢失分数,他们的排名也没有发生变化。
首先根据要求查询出未更新前的排名
SELECT *, row_number() over (order by points desc, name) AS `rank`
FROM TeamPoints
| team_id | name | points | rank |
| ------- | ----------- | ------ | ---- |
| 1 | Senegal | 2132 | 1 |
| 4 | Croatia | 1817 | 2 |
| 3 | Algeria | 1431 | 3 |
| 2 | New Zealand | 1402 | 4 |
其次查询出更新后的排名
SELECT one.team_id,
one.name,
(one.points + two.points_change) AS points,
row_number() over (order by (one.points + two.points_change) desc, one.name) AS `rank`
FROM TeamPoints one
LEFT JOIN PointsChange two USING(team_id)
| team_id | name | points | rank |
| ------- | ----------- | ------ | ---- |
| 1 | Senegal | 2110 | 1 |
| 3 | Algeria | 1830 | 2 |
| 4 | Croatia | 1830 | 3 |
| 2 | New Zealand | 1402 | 4 |
最后将两个SQL根据 team_id 进行连表,rank 相减即可(里面需要注意的点就是相减错误问题)
MySQL 当两个字段想减时,如果其中 一个或两个字段
的类型的 unsigned
无签名类型,
如果 差值
小于 0则会报错(BIGINT UNSIGNED value is out of range)
测试:
大于等于 0 的情况:
select cast(1 as unsigned) - 1
# 输入:0
select 2 - cast(1 as unsigned)
# 输入:1
select cast(2 as unsigned) - cast(1 as unsigned)
# 输入:1
小于 0 的情况:
select cast(1 as unsigned) - 2
# 报错:BIGINT UNSIGNED value is out of range in '(cast(1 as unsigned) - 2)'
select 0 - cast(1 as unsigned)
# 报错:BIGINT UNSIGNED value is out of range in '(0 - cast(1 as unsigned))'
select cast(1 as unsigned) - cast(2 as unsigned)
# 报错:BIGINT UNSIGNED value is out of range in '(cast(1 as unsigned) - cast(2 as unsigned))'
小于 0 的报错处理:
select cast(1 as signed) - 2
# 输出:-1
select 1 - cast(2 as signed)
# 输入:-1
select cast(1 as signed) - 2
或 select 1 - cast(2 as signed)
即可参考:处理 unsigned 相减错误(BIGINT UNSIGNED value is out of range)
SELECT a.team_id,
a.name,
(CAST(a.`rank` AS SIGNED) - CAST(b.`rank` AS SIGNED)) AS rank_diff
FROM (
SELECT *, row_number() over (order by points desc, name) AS `rank`
FROM TeamPoints
) AS a
LEFT JOIN (
SELECT one.team_id,
one.name,
(one.points + two.points_change) AS points,
row_number() over (order by (one.points + two.points_change) desc, one.name) AS `rank`
FROM TeamPoints one
LEFT JOIN PointsChange two USING(team_id)
) AS b USING(team_id)