问题的核心可以这样解决。。。

DROP TABLE IF EXISTS my_table;

CREATE TABLE my_table
(id SERIAL PRIMARY KEY
,timestamp int NOT NULL
,worker char(38) NOT NULL
);

INSERT INTO my_table VALUES
(1,  1351508000,     'worker1'), 
(2,   1351508020,     'worker1'),  
(3,   1351508060,     'worker2'), 
(4,   1351508070,     'worker1'),  
(5, 1351508089,     'worker1'), 
(6,   1351508200,     'worker1'),  
(7,   1351508300,     'worker1'), 
(8,   1351508370,     'worker2'),   
(9,   1351508410,     'worker2'), 
(10,  1351508450,     'worker2'),   
(11,  1351508530,     'worker1'), 
(12,  1351508540,     'worker2'),   
(13,  1351508580,     'worker2'), 
(14,  1351508600,     'worker2'),  
(15,  1351509290,     'worker2'), 
(16,  1351509380,     'worker1'),  
(17,  1351509390,     'worker1'),
(18,  1351509460,     'worker2'),  
(19,  1351509480,     'worker2'),  
(20,  1351509710,     'worker1'), 
(21,  1361510000,     'worker1'), 
(22,  1361510020,     'worker1'),   
(23,  1361510060,     'worker1'), 
(24,  1361510070,     'worker1') ;

SELECT x.*
     , CASE WHEN @prev=worker THEN @i:=@i ELSE @i:=@i+1 END i
     , @prev:=worker
  FROM my_table x
     , (SELECT @prev:=null,@i:=0) vars
 ORDER 
    BY timestamp;
+----+------------+---------+------+---------------+
| id | timestamp  | worker  | i    | @prev:=worker |
+----+------------+---------+------+---------------+
|  1 | 1351508000 | worker1 |    1 | worker1       |
|  2 | 1351508020 | worker1 |    1 | worker1       |
|  3 | 1351508060 | worker2 |    2 | worker2       |
|  4 | 1351508070 | worker1 |    3 | worker1       |
|  5 | 1351508089 | worker1 |    3 | worker1       |
|  6 | 1351508200 | worker1 |    3 | worker1       |
|  7 | 1351508300 | worker1 |    3 | worker1       |
|  8 | 1351508370 | worker2 |    4 | worker2       |
|  9 | 1351508410 | worker2 |    4 | worker2       |
| 10 | 1351508450 | worker2 |    4 | worker2       |
| 11 | 1351508530 | worker1 |    5 | worker1       |
| 12 | 1351508540 | worker2 |    6 | worker2       |
| 13 | 1351508580 | worker2 |    6 | worker2       |
| 14 | 1351508600 | worker2 |    6 | worker2       |
| 15 | 1351509290 | worker2 |    6 | worker2       |
| 16 | 1351509380 | worker1 |    7 | worker1       |
| 17 | 1351509390 | worker1 |    7 | worker1       |
| 18 | 1351509460 | worker2 |    8 | worker2       |
| 19 | 1351509480 | worker2 |    8 | worker2       |
| 20 | 1351509710 | worker1 |    9 | worker1       |
| 21 | 1361510000 | worker1 |    9 | worker1       |
| 22 | 1361510020 | worker1 |    9 | worker1       |
| 23 | 1361510060 | worker1 |    9 | worker1       |
| 24 | 1361510070 | worker1 |    9 | worker1       |
+----+------------+---------+------+---------------+

从这里开始，您可以在groupwise max（或min）的标准模式中使用上面的子查询，如下所示。。。

SELECT a.*
  FROM my_table a
  JOIN 
     ( SELECT MIN(id) id 
         FROM
            ( SELECT x.*
                   , CASE WHEN @prev=worker THEN @i:=@i ELSE @i:=@i+1 END i
                   , @prev:=worker
                FROM my_table x
                   , (SELECT @prev:=null,@i:=0) vars
               ORDER 
                  BY timestamp
            ) n
        GROUP 
           BY i
     ) b
    ON b.id = a.id;

+----+------------+---------+
| id | timestamp  | worker  |
+----+------------+---------+
|  1 | 1351508000 | worker1 |
|  3 | 1351508060 | worker2 |
|  4 | 1351508070 | worker1 |
|  8 | 1351508370 | worker2 |
| 11 | 1351508530 | worker1 |
| 12 | 1351508540 | worker2 |
| 16 | 1351509380 | worker1 |
| 18 | 1351509460 | worker2 |
| 20 | 1351509710 | worker1 |
+----+------------+---------+