If table reorganization is an option for you, you could consider using
integer ranges for describing parant/children relationships.
I have seen them in one of Joe Celko books:
http://www.amazon.com/gp/product/1558609202/sr=8-2/qid=1154595988/ref=pd_bbs_2/104-2243146-1376759?ie=UTF8
In that way,
On 8/2/06, bruce <[EMAIL PROTECTED]> wrote:
hi...
i have a tbl
fooTBL
name
parentID
ID
so a name can have might have a parentID, as well as an ID. 'name's are
associated with other 'name's via the parentID. in other words, if a name's
parentID == a name's ID, name1 is the parent of na
Typically, a temp table is used for this.
Beware of circular references.
Fred -> Wilma -> Betty -> Barney -> Fred
> -Original Message-
> From: [EMAIL PROTECTED] [mailto:pgsql-general-
> [EMAIL PROTECTED] On Behalf Of bruce
> Sent: Wednesday, August 02, 2006 11:55 AM
> To: 'PgSQL-General'