While I agree this is a reliable way of matching, this has a
significant impact on performance to make an extra query on each field
that I want the type of ! And this makes this approach definitly not
the one to implement IMO.
On 8/5/06, Volkan YAZICI <yazicivo(at)ttnet(dot)net(dot)tr> wrote:
> On Aug 04 04:38, Andro wrote:
> > Oids are in pg_type catalog (server side) and src/include/catalog/pg_type.h
> > (hard-wired).
> > But what should we compare the Oid returned by PQftype() with?
> > Let's say I want to check if column 1 is a VARCHAR, do I have to
> > if (PQftype(res,1) == 1043)
> > ...;
> > ?
> > Isn't there a kind of enum which we could rely on to find out types? What if
> > Oids change in pg_type.h?
> You can make another query on pg_type to learn OID of a specific type.
> And then compare it with the returned one from PQftype(main_query_res).
> Furthermore, I think this is the only reliable way of accomplising this
> task. (IIRC, PHP PostgreSQL API does same too.) Otherwise, when you use
> hard-wired headers (e.g. by including pg_type.h) what will you do when
> related server has a custom type that isn't get shippied with the
> PostgreSQL headers you included.
In response to
pgsql-interfaces by date
|Next:||From: Georgi Kolev||Date: 2006-08-09 11:38:24|
|Subject: PQoidValue - get last ID of primary key after INSERT - small fix|
|Previous:||From: Peter Eisentraut||Date: 2006-08-06 20:26:28|
|Subject: Re: getting the correct column size....|