Question

如果给定id有多个记录，则使用group by，having子句将告诉我。反正知道这两条记录在其余列中是如何相互不同的吗？

mysql>select pid, name, city from table1;
+------+-------------+--------+
| pid  | name        | city   |
+------+-------------+--------+
|    1 | aa          | delhi  | 
|    2 | bb          | delhi  | 
|    3 | cc          | mumbai | 
|    4 | salman      | pune   | 
|    4 | salman khan | pune   | 
+------+-------------+--------+
5 rows in set (0.00 sec)

mysql>select pid, count(*) as cnt from table1 group by pid having cnt > 1;
+------+-----+
| pid  | cnt |
+------+-----+
|    4 |   2 | 
+------+-----+
1 row in set (0.00 sec)

预期结果：

+------+-------------+
| pid  | name        |
+------+-------------+
|    4 | salman      | 
|    4 | salman khan | 
+------+-------------+
2 rows in set (0.00 sec)

我可以通过使用以下查询来实现这一目标......

mysql>select pid, name from table1 where pid=4;

但我怎么知道这两行名称不同而且城市是一样的呢？表中有一个时间戳列，我需要根据该时间对这些行进行排序。给定PID的Earlist记录将是第一个。

Answer 1

要获得您发布的预期结果，请尝试：

select pid, name 
from table1 
where pid in 
  (select pid
  from table1 
  group by pid 
  having count(*) > 1)
group by pid, name

如果您对

的情况特别感兴趣

城市和pid是相同的
名称不同
按每组内的时间排序

正如您在问题中解释的那样，请尝试：

select pid, name, city, timestamp 
from table1 
where pid in 
  (select pid
  from table1 
  group by pid, city 
  having count(*) > 1)
group by pid, name, city
order by pid, city, timestamp

Answer 2

通过Google搜索我发现Kodyaz.com中的以下示例可能会让您走上正确的轨道。

查找列值的差异

2 个答案: