on@uta.fi (Ossi Numminen) (10/24/90)
Dear Ingres gurus, One of our software vendors claims that the existence of views as such make base table updates more slowly. In principle this should not be the case. Is this really true with Ingres? If so, any estimates how much does it affect? ossi
bg0l+@andrew.cmu.edu (Bruce E. Golightly) (10/27/90)
That might be true, depending on what makes up the view and what objects are associated with the base table(s). A complex view consisting of a number of tables linked together might take a while to update. This could be complicated by base tables with inappropriate structures or lots of secondary indices.I suspect, though, that there is no appreciable difference between operating on a view and performing the same set of operations on the corresponding set of tables. How about more details?
on@uta.fi (Ossi Numminen) (10/29/90)
In article <wb_9t0_00WBNQ2jHdN@andrew.cmu.edu> bg0l+@andrew.cmu.edu (Bruce E. Golightly) writes: >That might be true, depending on what makes up the view and what objects >are associated with the base table(s). A complex view consisting of a >number of tables linked together might take a while to update. This >could be complicated by base tables with inappropriate structures or >lots of secondary >indices.I suspect, though, that there is no appreciable difference >between operating on a view and performing the same set of operations on >the corresponding set of tables. It is understandable that updating through a view takes some more time than just updating a base table directly. But if you have some views for just retrieving the information and then you update that under- lying base table directly, I see no reason why this update should take any longer than updating the same table in the case it had no views. And this is what they are claiming. > >How about more details? I'll try to get some. ---ossi