On Wed, 5 Apr 2006, Eugene E. wrote:
> Stephan Szabo wrote:
> > On Tue, 4 Apr 2006, Eugene E. wrote:
> >
> >
> >>Stephan Szabo wrote:
> >>
> >>>On Fri, 31 Mar 2006, Eugene E. wrote:
> >>>
> >>>
> >>>
> >>>>Peter Eisentraut wrote:
> >>>>
> >>>>
> >>>>>Eugene E. wrote:
> >>>>>
> >>>>>
> >>>>>
> >>>>>>the problem is: you'll get this four byte sequence '\000' _instead_
> >>>>>>of NUL-byte anyway.
> >>>>>
> >>>>>
> >>>>>What you seem to be missing is that PostgreSQL data can be represented
> >>>>>in textual and in binary form. What you in psql is the textual form.
> >>>>>If you want the binary form you need to select it. Then you can pass
> >>>>>the exact bytes back and forth.
> >>>>
> >>>>your sentence is not true.
> >>>>I can not select exact bytes even if i use BYTEA type
> >>>
> >>>
> >>>No, that is still using the textual form. If you use PQexecParams and set
> >>>the last argument to show you want binary data, you should get binary
> >>>data.
>
> ok
> then i am using PQexecParams
>
> the following tiny program shows a wonderful lameness...
What lameness? The fact that you're trying to use a binary (network order
maybe) integer as a string? That's not the fault of PQexecParams but
of the code calling it.
It'd be nice for ease of use to be able to say, give me this
column (the integer) as a string and this column (the bytea) as binary,
especially as the commentary implies that the protocol supports it. But as
I said before, that's a separate discussion from whether or not Peter's
claim that it's possible to get binary data is false.