Re: [gnome-db] gda_data_model_update_row syntax



> I don't really like this way, even if it should work in all cases.

Neither do I :-).


> For example I dislike when the wrong row will be updated in a
> homogeneous data model.

What's your definition of 'wrong' row?  Do you disagree with my
footnote [2] in the e-mail you replied to?  How?


> Anyway, drawback of this method is that we have to store all values of
> the updatable data model somewhere, in order to use them in the UPDATE
> instruction (or maybe I'm missing something).

You're right that we need to store the original values.  But we need
this only for rows requested by gda_data_model_get_updatable_row, not
the whole table.  This is only a small drawback, considering that in
some of our current provider implementations (e.g. mysql), two copies
of the whole table exist in the client memory (one held by mysql api,
one held by libgda).

If we really care about this drawback.  We can choose to implement
updatable model only in queries that contain the all primary keys (or
a unique non-null column) of a table.  This is what MS Access does.


> Nevertheless I will have a deeper look into this problem.  I hope to
> find a more elegant solution ;-)

The more people we have thinking about it, the better it gets :-)

Paisa



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]