Databases Reference
In-Depth Information
2.
The compensatory actions from S to ST and TC will happen automatically, precisely
because ST and TC are views of S; that is, updates to S will automatically be reflected
appropriately in ST or TC or both.
3.
The compensatory actions from ST and TC to S—these are the ones that are generally
thought of as the view updating rules as such—will also happen automatically, again
precisely because ST and TC are views of S.
13
That is, updates to ST and/or TC are
“really” updates to the underlying relvar S, and so are automatically visible in S, as well as
in ST or TC or both.
4.
A user who sees only views ST and TC can behave exactly as if those views were base
relvars (to the extent described previously, that is, no more and no less, for the case in
which they actually were base relvars).
Note:
There's a point here you might find a trifle odd, however, regarding explicit
UPDATE operations in particular. I've said that attempts to change a supplier's status in
relvar ST will either fail or have undesirable side effects (in general), and hence that such
operations probably shouldn't be requested in the first place. Under the conventional
approach to view updating, however (such as it is), such operations don't seem to be
problematic at all; that is, they seem to work as intuitively expected.
But that's because the
conventional approach tacitly assumes information equivalence
. That is, in the case at
hand, the conventional approach effectively assumes we're dealing with a design consisting
of relvars ST and
SC
(as in Example 1, earlier in the chapter), not a design consisting of
relvars ST and
TC
as in the present discussion. In Example 1, as you'll recall, attempts to
change a supplier's status in relvar ST worked just fine.
5.
By contrast, a user who sees just view ST or just view TC will clearly be limited in the
operations he or she is allowed to perform—again exactly as previously described in the
present section.
13
Actually there aren't any compensatory actions for relvar TC, but of course that fact doesn't invalidate the general point being
made here.
