Re: ecpg weird behavior

On Tue, Mar 19, 2002 at 07:46:04PM -0800, Thomas Lockhart wrote:
> No (although I was not aware that the above would work). The example
> looks like
> 
> exec sql begin declare section;
>   char *name_arr[10];
> exec sql end declare section;
> exec sql declare names cursor for
>   select name from horses;
> strcpy(msg, "open");
> exec sql open names;
> exec sql fetch names into :name_arr;
> exec sql close names;

This would mean the fetch knows how many tuples to return without you
telling it. 

How about:

...
exec sql open names;
exec sql fetch 10 in names into :name_arr;
exec sql close names;
...

The biggest plus of this way to write the statement is that it works
with ecpg. :-)

I just added this to my test cases. However, I didn't have to change
anything in ecpg sources.

>   Each FETCH returns, at most, the number of rows in the array
> dimension.

This has to be added. In fact ECPG has to translate the simple fetch
into a fetch <how_many> ... since otherwise the backend wouldn't return
more than one tuple.

>   Fewer rows are returned in the following cases: <snip cases>
>   The cumulative number of rows returned can be found in the third
> element
>    of sqlerrd in the SQLCA, called sqlerrd[2] in this guide.

This works with ECPG as well.

> I'm guessing that this is a relatively short hop from your existing
> array capabilities, but don't how close. What do you think?

Quite easy I guess. Needs to be added to the TODO list. :-)

Michael

-- 
Michael Meskes
Michael@Fam-Meskes.De
Go SF 49ers! Go Rhein Fire!
Use Debian GNU/Linux! Use PostgreSQL!