The Origin Forum - xyz_renka broken ?

The Origin Forum

Username:	Password:
Save Password
Forgot your Password? \| Admin Options

All Forums

Origin Forum for Programming

LabTalk Forum

xyz_renka broken ?

New Topic

Reply to Topic

Printer Friendly

Author

Topic

HaroldPit

3 Posts

Posted - 11/24/2011 : 11:27:40 AM

Origin Ver. and Service Release: Origin Pro 8.5.0G SR1
Operating System: WinXP

Hi everyone,

I have a problem with the xyz_renka-function when it's called from the "command line". I'm sure it worked at some point in my scripting-endeavour and it still works in projects that I haven't worked in before so I likely just screwed something up. Unfortunately I cannot figure out, what went wrong and how to fix it.

Whenever I use the xyz_renka function now it returns a matrix with only "--" in the uppermost rows(depending on the total amount of rows) and very small values for the rows at the bottom (something like "1.23456E-200").

I tried the following commands

xyz_renka 3;
xyz_renka 3 rows:=5 cols:=5;

Where the last one returns this:

--		--		--		--		--
--		--		--		--		--
--		--		--		--		--
4,22576E-202	6,02098E-200	1,19997E-199	1,79784E-199	2,39572E-199
1,6903E-201	2,40839E-199	4,79988E-199	7,19137E-199	9,58286E-199

However when I use the XYZ-Gridding Assistant it works just fine (altough it fails to recognize the correct number of rows).

Maybe I assigned a variable, that is used by the gridding function but I cannot figure out what went wrong. I would be glad for any hint what to look into.

Penn

China
644 Posts

Posted - 11/29/2011 : 04:17:40 AM

Hi,

Please send your data or project to Technical Support. Then we can have a try with your data to see whether we can reproduce your problem. Please refer to this post in your email.

Penn

greg

USA
1379 Posts

Posted - 12/01/2011 : 11:38:26 AM

Some of the 'raw' gridding functions (like xyz_renka) are not really useful since they rely on other internal parameters having been set by the 'Convert Worksheet to Matrix' dialog, which has already pre-analysed your data and determined suitable conversion method, columns, rows, etc. when it opens.

Since 8.5 we have provided the XYZ2MAT and XYZ2MATLOG X-Functions to leverage the pre-analysis used in the dialog.

In your case, you can use
XYZ2MAT 3 settings:ConvertToMatrix:=2; // Here, 2 is Renka_Cline

For all methods:
0 Regular
1 Sparse
2 Renka-Cline
3 Shepard
4 Thin Plate Spline
5 Kriging Correlation
6 Weighted Average

You can also control data examination and duplicate handling:

XYZ2Mat iz:=[Book1]Sheet1!(A,B,C)
settings.ExamData.XPrecision:=1E-8
settings.ExamData.YPrecision:=1E-8
settings.ExamData.RemoveDuplicate:=0
settings.ConvertToMatrix:=2
om:=<new>;

For Duplicate handling:
0 Mean
1 Median
2 Minimum
3 Maximum
4 Sum
5 None

I notice your data is really regular data with some noise in the XY coordinates and some missing values. Consequently,
xyz_regular 3;
will work for you even though you will get a message about it not being regular (due to tolerance).
The best method (although slow) for your data (which is very noisy) is Thin Plate Spline:
xyz_tps 3;
run.section(Plot3D,ColorMap);

Edited by - greg on 12/01/2011 11:52:38 AM

HaroldPit2

2 Posts

Posted - 12/01/2011 : 3:31:08 PM

Awesome, thanks for the help Greg. I can't wait to try it tomorrow.

quote:
which has already pre-analysed your data

I was suspecting something along these lines. Those nasty helpful blackboxes.

FYI: I lost my login credentials so I had to register a new count.

Topic

New Topic

Reply to Topic

Printer Friendly

Jump To:

The Origin Forum

Snitz Forums 2000