Martin Kutschker writes:
 > Have the systems which use sequences some nameing schemes, which build the 
 > sequence name from the table name?
 > 

SQLite and MySQL don't use sequences, so I can speak only of
PostgreSQL (don't know about Oracle, Sybase et al.). If you use the
Serial type in PostgreSQL, a sequence is automatically created with a
predictable name (tablename_colname_seq). However, you can do this
just as well:

CREATE SEQUENCE whatever_seq;
CREATE TABLE tablename (
    colname integer DEFAULT nextval('whatever_seq') NOT NULL

A mechanism that relies on a particular naming scheme of sequences
would fail in the second case.

 > And perhaps dbi couild handle the problem that Mysql doesn't store the id by 
 > table or sequence but by connection by automatically calling 
 > mysql_last_insert_id() on every insert and storing the ids in a dbi 
 > datastructure per table.
 > 

I don't see a real problem here. If an application makes sure to store
relevant id values after INSERT commands, you'll get a portable code
without causing dbi to interfere with tables behind the users back.

regards,
Markus

-- 
Markus Hoenicka
[EMAIL PROTECTED]
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de



-------------------------------------------------------
All the advantages of Linux Managed Hosting--Without the Cost and Risk!
Fully trained technicians. The highest number of Red Hat certifications in
the hosting industry. Fanatical Support. Click to learn more
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=107521&bid=248729&dat=121642
_______________________________________________
libdbi-users mailing list
libdbi-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libdbi-users

Reply via email to