just want to make sure I understand the convention for libpq and libpqxx.
the convention for getting data into postgres is to convert everything
to strings first? even if the data is already stored as a 64bit
double in the client app?
On Thu, Dec 11, 2008 at 11:20 AM, Jeroen Vermeulen <jtv(at)xs4all(dot)nl> wrote:
> Whit Armstrong wrote:
>> can copy be used to append data to a table?
> Yes. The only thing it will not do for you is update existing rows.
>> The only reason I'm hesitant to use the copy command is because I
>> would have to render every row of my data to a string, and 64bit
>> doubles could mean long strings, formating issues, etc.
>> roughly what I was thinking was (this is pseudocode):
>> PGtuple myTuple = createTuple();
>> PGinsert_tuple(conn, myDestinationTable, myTuple);
>> are there facilities to do something like that, or am I really reaching
> In libpqxx I mostly provide self-made text conversions, but you can also use
> the standard library to "print" most values into text buffers. Just be sure
> to use the C locale where you do this, or you might end up with strange bugs
> depending on local notations!
In response to
pgsql-interfaces by date
|Next:||From: Bosco Rama||Date: 2008-12-12 05:22:03|
|Subject: gcc _Decimal types & ecpg|
|Previous:||From: Jeroen Vermeulen||Date: 2008-12-11 16:20:55|
|Subject: Re: simple example of copying data from a c/c++ array