One of the books you recommend, Introduction and concepts,
specifically notes using OID as a unique for primary and foreign keys.
Pg 80.
If there is a danger, Id like to know what it is, since I am using
the OID as a unique in a join to generate a list of records for
individual processing.
As long as you remember that you cannot set, or alter an OID
what's the diff?
L>
At 8:21 AM -0800 4/5/02, Josh Berkus wrote:
>Harald, Chris,
>
>> > 1. M$ SQL Server has an identity column which generates a
>> sequential number
>> > for you, we use this often for unique primary keys. Does postgres
>> support
>> > anything like this ?
>>
>> select *,oid from your_table
>
>NO!!!! This is wrong. DO NOT USE the oid for anything other than
> system purposes. It is NOT a user-friendly value for several reasons.
>
>Chris, the data type you want is SERIAL. Please read the section of
> the docs on sequences:
>http://www.postgresql.org/idocs/index.php?sql-createsequence.html
>... as the SERIAL datatype simply automates the creation of a sequence.
> BTW, postgres sequences are both more powerful and more user-friendly
> than MS SQL Identity columns.
>
>I would also strongly suggest buying a beginner's postgresql book, such
> as "PostgreSQL: Introduction and Concepts" or Wrox Press' PostgreSQL
> book.
>
>-Josh Berkus
>
>---------------------------(end of broadcast)---------------------------
>TIP 6: Have you searched our list archives?
>
>http://archives.postgresql.org
--
searchmaster@apple.com
lev@apple.com
