[KLUG Members] Sequences & MySQL

[Jamie McCarthy]

awilliam at whitemice.org (Adam Tauno Williams) passes along the
question:

> As far as I can see the only option is to create a helper table
> which contains one otherwise unused record for each unique key.
> Creating a new key would imply inserting a record in the helper
> and retrieving the auto-increment value (which is then used in
> the real tables).
>
> Obviously this is kinda fragile, so maybe someone knows a better
> workaround? :-|

That's the recommended way to implement sequences:

http://dev.mysql.com/doc/mysql/en/information-functions.html

(scroll down to "LAST_INSERT_ID")

In practice, there's no reason for the table that holds the sequence
counter to be empty:  one may as well put real data into it, too.

I'm afraid I don't understand how this is any more "fragile" than
any other way...
-- 
  Jamie McCarthy
 http://mccarthy.vg/
  jamie at mccarthy.vg