Re: [gnome-db] gda_connection_execute_non_select_command ()

[Bas Driessen <bas driessen xobas com>]

On Mon, 2006-09-11 at 16:11 +0200, Vivien Malerba wrote:
> On 9/11/06, Bas Driessen <bas driessen xobas com> wrote:
> >
> >  Hi Vivien,
> >
> >  Testing the new gda_connection_execute_non_select_command
> > () function at the moment, but this function always returns -2. I have
> > tested for both MySQL as PostgreSQL. The data is updated correctly in the
> > database, but the return value is -2. Is this function still under
> > construction?
> >
> >  Example use:
> >
> >  GdaCommand *command = NULL;
> >  gint number = 0;
> >
> >  const gchar *sql = "update parts set price = 1.23 where part_id = '12345'"
> >
> >  command = gda_command_new (buffer, GDA_COMMAND_TYPE_SQL,
> > GDA_COMMAND_OPTION_STOP_ON_ERRORS);
> >
> >  number = gda_connection_execute_non_select_command
> > (connection, command, NULL, NULL);
> >
> >  printf("NUMBER !! %d\n", number)
> >
> >  gda_command_free (command);
> >
> >  Let me know if you see something obvious that I do wrong?
>
> No, you did nothing wrong. I did some code and as I was in a hurry I
> did not test it properly. Anyway, the problem is in the new
> gda_parameter_list_new_inline() function. Please apply the attached
> patch and it should be better.

Hi Vivien,

Did some initial testing for both PostgreSQL and MySQL. PostgreSQL works as expected. If I update 3 rows, I get the value 3 back. However, MySQL behaves differently. If I update 3 rows, it gives me back the number 3, which is correct. If I then execute the same command again, it returns 0. Reason is that even though I update the same 3 rows again, it really does not change the data in the end. Theoretically this is correct, but if I execute the same UPDATE statement in the mysql console multiple times, I always get the number 3 back.

So, both PostgreSQL as MySQL should behave the same I reckon. Since both the psql as mysql console always report the number of rows affected (even though data may physically not been changed), I believe we should go with that approach and get MySQL results in line with PostgreSQL results.

Thanks,
Bas.