Re: importing db as text files

On Wed, Aug 13, 2003 at 20:45:55 -0700,
expect <expect(at)ihubbell(dot)com> wrote:
> On Wed, 13 Aug 2003 16:30:04 -0500
> Bruno Wolff III <bruno(at)wolff(dot)to> wrote:
>
> > On Wed, Aug 13, 2003 at 14:14:20 -0700,
> > expect <expect(at)ihubbell(dot)com> wrote:
> > >
> > > What's the big deal with importing text files? I have a 70 MB file to import
> > > and it's been one problem after another. I used the copy command and it appears
> > > that it's just not possible. I finally massaged the file into a .sql file and
> >
> > That is unlikely, but without more details it is hard to say what you need
> > to do.
>
>
> What's unlikely? That it didn't work? But it didn't work.

That it isn't possible to load your data using the copy command.
You may need to do some transformation before sending it to copy.
Using copy is probably going to give you the fastest load time.

> From what I've read it should use the default value when there is no value.
> Is that not true? Is this a known issue?

Where did you read that? What do you mean by 'no value'?

The syntax for the insert command is given in the manual:
INSERT INTO table [ ( column [, ...] ) ]
{ DEFAULT VALUES | VALUES ( { expression | DEFAULT } [, ...] ) | query }

>
> No you've come to the wrong impression. I believe that the problem lies in the
> fact that pg will not default to the default value when no value is present.

And how do you think 'no value' is represented? If you use default to
respresent 'no value' then things will work as you expect for insert
statements. When you are using copy, defaults can't be used on a row
by row basis (but unlisted columns will get default values) and you will
need to inlcude the value that you want in each row.