home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
tsql
/
doc
/
tdbglossary.mail
/
000016_ahn@cbnmva.att.com _Thu Sep 16 07:35:21 1993.msg
< prev
next >
Wrap
Internet Message Format
|
1996-01-31
|
3KB
Received: from univers.CS.Arizona.EDU by optima.CS.Arizona.EDU (5.65c/15) via SMTP
id AA19709; Thu, 16 Sep 1993 07:35:22 MST
Date: Thu, 16 Sep 1993 07:35:21 MST
From: ahn@cbnmva.att.com
Message-Id: <199309161435.AA13705@univers.cs.arizona.edu>
Received: from att.UUCP by univers.cs.arizona.edu; Thu, 16 Sep 1993 07:35:21 MST
To: gadia@cs.iastate.edu
Cc: tdbglossary@cs.arizona.edu
Subject: Re: Ahn/Temporal element
From: Shashi K. Gadia <gadia@cs.iastate.edu>
Subject: Re: Ahn/Temporal element
To: ahn@cbnmva.att.com
Date: Wed, 15 Sep 93 14:36:27 CDT
Cc: tdbglossary@cs.arizona.edu
Gadia,
This discussion is more interesting and complex than I initially thought.
I still have some questions.
> If a union is a set operator, doesn't a union of intervals
> result in a set (of intervals) ?
No.
Is the union not a set operator then ?
This distinction is similar to: "a relation is not a sequence
of tuples rather a set of tuples. This distinction is
extremely important.
You may call a sequence a set, though the reverse is not true.
If the union is a set operator, then the result of a union may be
called a set, though the reverse is not true.
> Term "interval" is used in a general sense too.
> For example, Section 3.15 mentions bitemporal intervals.
>
What when it becomes a union of triangles?
Well, you may consider a triangle as a degenerate case of a rectangle.
(It may actually be a rectangle with two vertices 0.01 anstrom apart :)
Again, the reverse is not true.
... To me temporal
element (the concept) is the most primitive notion in
temporal databases.
I agree that it is an important concept.
The current proposal for TSQL2 is based on the concept.
That is why we need a good name that can withstand scrutinies of
people outside the temporal DB field.
A union of intervals is already a reduced object having
a unique representation, where as a set is not.
Rick wrote,
> This discussion raises two questions in my mind. If I have
> three temporal elements, each containing one interval:
> A = [1..10]
> B = [11..20]
> C = [21..30]
>
> Then D= A union C is a temporal element containing two intervals.
>
> My question is, how many intervals are in E = D union B?
Is the answer 1 or 3 ?
E = [1..10] U [21..30] U [11..30].
This looks like 3.
Can it be simplified (coalesced) to E = [1..30] ?
Then the answer is 1, which seems to me correct.
Example:
E = 1..10 U 5..23 U 21..30 U 6..24
Where as the above union is unique, the intervals chosen
to express that union is not.
You mean that E cannot be coalesced to 1..30 ?
Then what is the union operator for ?
Actually, it may be clearer to represent E in this case as a set,
without the union operator:
E = { 1..10, 5..23, 21..30, 6..24}
which is certainly different from
E = { 1..11, 5..23, 21..30, 6..24} or E = { 1..30 }
Regards,
Ilsoo