pg.ddx.io  pgsql-sql@postgresql.org mailing list archive  
help / color / mirror / Atom feed
database abstraction -> functions
2+ messages / 2 participants
[nested] [flat]

* database abstraction -> functions
@ 2002-09-24 07:45 Jeroen Olthof <j.olthof@millipede.nl>
  2002-09-25 14:03 ` Re: database abstraction -> functions Dmitry Tkach <dmitry@openratings.com>
  0 siblings, 1 reply; 2+ messages in thread

From: Jeroen Olthof @ 2002-09-24 07:45 UTC (permalink / raw)
  To: pgsql-general@postgresql.org; pgsql-sql

Hi,

When developing applications is a good thing to create abstraction between
different layers
The one concerning the database would be the persistence layer. To create
such abstraction I want all mij datababase activitie runned through
functions. But how can I return a set of rows instead of a single datatype
(I looked into the setof but never found a clear simple)?

kind regards
Jeroen Olthof







^ permalink  raw  reply  [nested|flat] 2+ messages in thread

* Re: database abstraction -> functions
  2002-09-24 07:45 database abstraction -> functions Jeroen Olthof <j.olthof@millipede.nl>
@ 2002-09-25 14:03 ` Dmitry Tkach <dmitry@openratings.com>
  0 siblings, 0 replies; 2+ messages in thread

From: Dmitry Tkach @ 2002-09-25 14:03 UTC (permalink / raw)
  To: Jeroen Olthof <j.olthof@millipede.nl>


Jeroen Olthof wrote:
> Hi,
> 
> When developing applications is a good thing to create abstraction between
> different layers
> The one concerning the database would be the persistence layer. To create
> such abstraction I want all mij datababase activitie runned through
> functions. But how can I return a set of rows instead of a single datatype
> (I looked into the setof but never found a clear simple)?
> 

A very simple answer - there is no way to do what you want :-(
They promise to have it in 7.3...

I am using a workaround, where my functions glue all the columns together into a pipe-separated
line, that I then parse back into columns in my application...

Also, note that, unless your functions are in C (anything like plpgsql, or just sql), they are slow like hell...
In that case, you are much better off implementing your abstraction layer through views...

Dima




^ permalink  raw  reply  [nested|flat] 2+ messages in thread


end of thread, other threads:[~2002-09-25 14:03 UTC | newest]

Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2002-09-24 07:45 database abstraction -> functions Jeroen Olthof <j.olthof@millipede.nl>
2002-09-25 14:03 ` Dmitry Tkach <dmitry@openratings.com>

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox