Bruno Wolff III wrote:
> On Tue, Jun 22, 2004 at 13:40:03 -0400,
> Matthew Nuzum <matt@followers.net> wrote:
>
>> The end result is to duplicate the data for a particular record in table
>> "a" so that all of it's related data in tables "b" and "c" is duplicated.
>>Where "b" is the middle table in a many to many relationship. For example,
>>
>>- a ---- - b ---- - c ----
>> aid <--+ bid <--+ cid
>> data1 +-->aid +-->bid
>> data2 field1 info1
>> date3 field2 info2
>>
>>SELECT a.*,b.*,c.* from a,b,c where b.aid = a.aid and c.bid = b.bid;
>>
>>So, the goal is to duplicate an object that is made up of the data stored
>>across these three tables.
[snip]
> Offhand I can't think of a way to avoid using a procedural language to walk
> through the b table so that you can easily keep track of which new bid
> corresponds to which old bid. However, writing a perl script or plsql
> function to do this for you shouldn't be difficult.
Screams out plpgsql to me - it's good at automating what is basically
cut & paste of values.
-- Richard Huxton Archonet Ltd
