先来创建一个步长为2的序列，最大值为10000，每次产生100个值。
t_girl=# create sequence ytt.ytt_s1 start with 1 increment by 2  maxvalue 10000 ;
CREATE SEQUENCE
创建一个测试表。
t_girl=# create unlogged table ytt.tmp_3 (id int not null, log_date date);                       
CREATE TABLE
改变表tmp_3的列id 默认值是序列ytt_s1的下一个值。
t_girl=# alter table tmp_3 alter id set default nextval(‘ytt_s1‘);
ALTER TABLE
t_girl=# \d tmp_3
                   Unlogged table "ytt.tmp_3"
  Column  |  Type   |                  Modifiers                   
----------+---------+----------------------------------------------
 id       | integer | not null default nextval(‘ytt_s1‘::regclass)
 log_date | date  


 OK，我们试着插入几条记录。
t_girl=# insert into tmp_3(log_date) select generate_series(‘2014-01-01‘::timestamp,now(),‘1 day‘);
INSERT 0 14
t_girl=# select * from tmp_3;
 id |  log_date  
----+------------
  1 | 2014-01-01
  3 | 2014-01-02
  5 | 2014-01-03
  7 | 2014-01-04
  9 | 2014-01-05
 11 | 2014-01-06
 13 | 2014-01-07
 15 | 2014-01-08
 17 | 2014-01-09
 19 | 2014-01-10
 21 | 2014-01-11
 23 | 2014-01-12
 25 | 2014-01-13
 27 | 2014-01-14
(14 rows)

我们先重置原来ID的默认值。
t_girl=# alter table tmp_3 alter id set default 0;
ALTER TABLE


清空测试表：
t_girl=# truncate table tmp_3;
TRUNCATE TABLE
现在的序列已经增长到了27，如下
t_girl=# \d ytt_s1;
     Sequence "ytt.ytt_s1"
    Column     |  Type   | Value  
---------------+---------+--------
 sequence_name | name    | ytt_s1
 last_value    | bigint  | 27
 start_value   | bigint  | 1
 increment_by  | bigint  | 2
 max_value     | bigint  | 10000
 min_value     | bigint  | 1
 cache_value   | bigint  | 1
 log_cnt       | bigint  | 19
 is_cycled     | boolean | f
 is_called     | boolean | t


 现在重置下序列。
 t_girl=# alter sequence ytt_s1 restart with 1;
ALTER SEQUENCE
t_girl=# \d ytt_s1
     Sequence "ytt.ytt_s1"
    Column     |  Type   | Value  
---------------+---------+--------
 sequence_name | name    | ytt_s1
 last_value    | bigint  | 1
 start_value   | bigint  | 1
 increment_by  | bigint  | 2
 max_value     | bigint  | 10000
 min_value     | bigint  | 1
 cache_value   | bigint  | 1
 log_cnt       | bigint  | 0
 is_cycled     | boolean | f
 is_called     | boolean | f


创建一个触发器函数。
create or replace function sp_increment_tmp_3()
returns trigger as
$ytt$
begin
  new.id := nextval(‘ytt_s1‘);
  return new;
end;
$ytt$ language plpgsql;


创建触发器。
create  trigger tr_insert_tmp_3 
before insert on ytt.tmp_3
for each row   
  execute procedure sp_increment_tmp_3();


  OK。 再次插入后，看看结果如何。
t_girl=# insert into tmp_3(log_date) select generate_series(‘2014-01-01‘::timestamp,now(),‘1 day‘);
INSERT 0 14
t_girl=# select * from tmp_3;
 id |  log_date  
----+------------
  1 | 2014-01-01
  3 | 2014-01-02
  5 | 2014-01-03
  7 | 2014-01-04
  9 | 2014-01-05
 11 | 2014-01-06
 13 | 2014-01-07
 15 | 2014-01-08
 17 | 2014-01-09
 19 | 2014-01-10
 21 | 2014-01-11
 23 | 2014-01-12
 25 | 2014-01-13
 27 | 2014-01-14
(14 rows)

现在呢，删掉之前的触发器和触发器函数。然后创建一张新表。
看看下面，跟第一种是一样的，只不过是系统自己创建了一个序列。
 t_girl=# create unlogged table ytt.tmp_3 (id smallserial not null, log_date date);           
CREATE TABLE
t_girl=# \d tmp_3;
                        Unlogged table "ytt.tmp_3"
  Column  |   Type   |                     Modifiers                      
----------+----------+----------------------------------------------------
 id       | smallint | not null default nextval(‘tmp_3_id_seq‘::regclass)
 log_date | date     | 


t_girl=# \d tmp_3_id_seq
          Sequence "ytt.tmp_3_id_seq"
    Column     |  Type   |        Value        
---------------+---------+---------------------
 sequence_name | name    | tmp_3_id_seq
 last_value    | bigint  | 1
 start_value   | bigint  | 1
 increment_by  | bigint  | 1
 max_value     | bigint  | 9223372036854775807
 min_value     | bigint  | 1
 cache_value   | bigint  | 1
 log_cnt       | bigint  | 0
 is_cycled     | boolean | f
 is_called     | boolean | f
Owned by: ytt.tmp_3.id

PostgreSQL 实现MySQL 的auto_increment 字段