I've looked at your file - thanks for sending it along. You have 40,000
points using a MathType (x,y)->z, with one sample per line in the text
file. This will cause the creation of an Irregular2DSet of 40,000
points for the domain (since this form implies irregularly-spaced data)
and the Delaunay algorithm on 40K 2D points will, indeed, be slow as it
tries to make connections between the points.
On my computer, it took over 11 minutes to perform these computations
(the data were read from disk in just a couple of seconds).
I did not do a thorough analysis of your data, but it appeared that the
sampling set might essentially be a 2D matrix (grid) of equally-spaced
points. If that is true, and you will be processing a lot of these
data, I would recommend that you alter your input text file format so
that you can use:
That is, put the 200 values of for x(0:39) (the increment for the
sampling x coordinate will be computed) on each line (row), and have 200
lines (rows) in the file. This will reduce your 'reading' time to just
a few seconds.
If, on the other hand, you do not need such a topology, the you could
use a MathType of (index -> (x,y,z)), in which case the 'definition
lines' in your file would read:
On Wed, 28 Aug 2002, yahya ay wrote:
> when i run my program i get that :
> Specified MathType = ((x,y)->(z))
> from: ((x,y)->(z)) made MathType: ((x, y) -> z)
> and it takes 6 min to resd data from a file.
> is that ok???????
University of Wisconsin-Madison
Space Science and Engineering Center
Cooperative Institute for Meteorological Satellite Studies