PostgreSQL の timestamp with time zone の等価判定について ref: http://qiita.com/a-suenami/items/dea19256f6ec672d67f1

file0.txt

postgres=# CREATE TABLE orders ( product_id integer, ordered_at timestamp with time zone, CONSTRAINT pkey PRIMARY KEY (product_id, ordered_at) ) ;
CREATE TABLE

postgres=# \d orders 
               Table "public.orders"
   Column   |           Type           | Modifiers 
------------+--------------------------+-----------
 product_id | integer                  | not null
 ordered_at | timestamp with time zone | not null
Indexes:
    "pkey" PRIMARY KEY, btree (product_id, ordered_at)

file1.txt

postgres=# INSERT INTO orders (product_id, ordered_at) VALUES (1, '2014-07-01 14:00:00') ;
INSERT 0 1

file2.txt

postgres=# SELECT product_id, ordered_at FROM orders ;
 product_id |       ordered_at       
------------+------------------------
          1 | 2014-07-01 14:00:00+09
(1 row)

file3.txt

postgres=# INSERT INTO orders (product_id, ordered_at) VALUES (1, '2014-07-01 14:00:00 UTC') ;
INSERT 0 1
postgres=# SELECT product_id, ordered_at FROM orders ;
 product_id |       ordered_at       
------------+------------------------
          1 | 2014-07-01 14:00:00+09
          1 | 2014-07-01 23:00:00+09
(2 rows)

file4.txt

postgres=# INSERT INTO orders (product_id, ordered_at) VALUES (1, '2014-07-01 05:00:00 UTC') ;
ERROR:  duplicate key value violates unique constraint "pkey"
DETAIL:  Key (product_id, ordered_at)=(1, 2014-07-01 14:00:00+09) already exists.