[GRASSLIST:813] Re: postgres connection

[mdeneen at japple.org]

On Fri, 25 Jul 2003, Radim Blazek wrote:

> On Friday 25 July 2003 16:32, mdeneen at japple.org wrote:
> > The reason I want to go straight to postgres is because when I try to
> > import using dbf, the dbf process grows up to 1000 megs before I kill it.
> > I don't know if I have enough memory to import the CompleteChain layer.
> >
> > Is this something that will definately require a lot of memory?
> >
> > -M
> >
> > On Fri, 25 Jul 2003 mdeneen at japple.org wrote:
> > > How do I set up grass 5.1 so that v.in.ogr will import the data into a
> > > postgres table?  I have found several tutorials, but they all deal with
> > > taking existing grass data and inserting it into a database.
> > >
> > > I'm sure it's simple, but I couldn't find the grass 5.1 documentation for
> > > it.
> > >
> > > -M
> 
> You can define driver/database by v.database driver=name database=name,
> but the database must exist before (it is not created by GRASS).
> 
> Dbf driver creates table in the memory and writes it to disk when driver
> is closed, it may be also some memory leakage there. 
> Do you have a large layer? How many features? I worry that postgres 
> will be too slow for large data sets. 

It looks like that worked.  I am importing the CompleteChain layer from 
the tiger2k files, for the entire state of New Jersey, USA.

I also have some mapinfo tables, but they do not have as much information 
as the tiger2k files.

-M