mapi.connect / mapi_connect (Second Try)

Stephen P. Morgan stephen_p_morgan at
Mon Mar 11 16:56:40 CET 2013

In fact, then all of the sample programs, sample0.c--sample4.c and 
smack00.c--smack01,c pass NULL for the database field.

From: Stephen P. Morgan <stephen_p_morgan at>
To: Communication channel for developers of the MonetDB suite. 
<developers-list at>
Sent: Mon, March 11, 2013 8:53:15 AM
Subject: Re: mapi.connect / mapi_connect (Second Try)

Then the sample0 program is incorrect, as it passes NULL in the database field.

From: Fabian Groffen <fabian at>
To: developers-list at
Sent: Mon, March 11, 2013 8:35:18 AM
Subject: Re: mapi.connect / mapi_connect (Second Try)

On 11-03-2013 08:30:01 -0700, Stephen P. Morgan wrote:
> I wrote an email a few days ago to the effect that I could not get the
> mapi.connect()  MAL statement to work due to the server side complaining that
> I had not specified the database to use.
> This morning I ran one of the sample programs included with MonetDB,
> clients/examples,C/sample0.c (sample0).  It seems very similar as it issues
> a mapi_connect() with the database field set to NULL.  It fails in the same
> way, i.e.,
>     MAPI  = monetdb at localhost:50000
>     ACTION= mapi_start_talking
>     ERROR = !monetdbd: please specify a database
> Should mapi_connect() work with the database field set to NULL?  (If so,
> then the mapi.connect() MAL statement should work, I think.)  Is there a bug
> in the server that insists on the database being specified?  If there is a
> bug, does anyone have a fix for this?  It would be greatly appreciated.

No, you  should pass the database name, otherwise the server can't know
which one you'd like to connect to.

Fabian Groffen                              fabian at
column-store pioneer    
developers-list mailing list
developers-list at
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the developers-list mailing list