Stephen,
My initial guess is that it is to do with the way that MapInfo does its
intersection calcs. Like most GIS packages mapInfo probably uses an objects
Minimum Bounded Rectangle (MBR) to initially determine if a object is likely
to touch another object.
The mathematical comparison on two objects MBR'ss is straight forward and
quick to check. Once it is determined that two objects MBR's intersect then
a series of other calculations will be undertaken to determine if the
objects themselevs touch - this is obviously a much longer calculation
process.
What you have done by combining everything into one object is create an
object with a large MBR (and also dramatically increased the number of other
objects whose MBR's intersect the large MBR.)
In regard to your specific example where your are looking for an
intersection of one spefic object I would hazard a guess that the complexity
of the large shape makes the calculation more complex.
BTW I am not sure why you are concerned with the performance I have programs
that search for intersecting objects in recordsets that contain 200,000+
objects and have had no response time problems. If there is an issue
perhaps it is in the order of your calculations - I find that using the
'update column.....' command into a temporary column (then querying the
column) can be significantly quicker than moving through each record
individually.
Martin
==================================
Martin Roundill
GIS Manager
Waitakere City Council
Private Bag 93109
Henderson
Waitakere City
New Zealand
-----Original Message-----
From: Stephen R. Riese [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, 19 January 2000 08:09
To: MapInfo
Subject: MI One vs. Many Objects
Greetings,
I tried an experiment and the results are not what I expected. I'd like to
know if my results are an anomaly or should be expected.
I have a procedure that examines each object (about 200) in a table for
intersection with a given object. I tried this two ways:
1) Individual objects in the table are left as individual objects.
2) The objects are combined into one large object (and the table now has
only one row).
I expected the intersection detection routine to work faster in situation
number 2, where there is only one object. In my trials, there is a slight
improvement, but not much. Situation 1 actually is faster in some (few)
cases. Any thoughts on this and the inner workings of the "Obj1 Intersects
Obj2" function? I'm thinking that 200 objects in a table might not be
enough to produce a significant difference, but that if there were, say,
2000 or more objects I might see a difference.
Thanks,
Steve
[EMAIL PROTECTED]
----------------------------------------------------------------------
To unsubscribe from this list, send e-mail to [EMAIL PROTECTED] and put
"unsubscribe MAPINFO-L" in the message body, or contact [EMAIL PROTECTED]
----------------------------------------------------------------------
To unsubscribe from this list, send e-mail to [EMAIL PROTECTED] and put
"unsubscribe MAPINFO-L" in the message body, or contact [EMAIL PROTECTED]
