[opendx-users]timeseries 2-vector field generation
Hi, I made 2-vector field generation with alternative netCDF file. I haven't resolved the problem with the data from *.general import. By the way, I have another question. Is there any body who uses openDXto visualize the MM5 output? I have read the mailing archives but they were not enough for me.-.- I have changed MM5 output to netCDF format file with a MM5 utility and then converted the netCDF file to OpenDX-supported netCDF format file with ncdx(you know if?). After that I tried to import it to opendx but failed. OpenDX importer doesn't recognize the netCDF format. Because ncdx adds position attributes to variables with new dimensions which don't exist in netCDF file. I think ncdx does not work in this point for MM5 output.Thanks for any help. -Lee HanSooResearch Center for Disaster Environment, Disaster Prevention Research Institute,Kyoto University, Gokasho, Uji, Kyoto, Japan.611-0011 Tel: 0774-38-4309M.B : 080-5301-6649
Re: [opendx-users]timeseries 2-vector field generation
Be aware that Reduce will interpolate your data, which you probably don't want. That was just for testing with a lower-res grid. Also be sure that Include's min and max args are vectors since you are feeding vector data, else it will take the mag(vector) if you send min-max as scalars. If there are no invalids, do you have any problem with the vectors? You could test this by Compute( a == - ? 0 : a ) to temporarily override your invalids. Be assured that under normal circumstances, what you propose to do works. The vectors should emit from their positions (which are treated as local origins for each vector) and extend in their vectorial direction (given by the data). The only thing that should change is that the invalidated ones will disappear. If you try the substitution idea, the invalids will become spheres at their base positions (but of course, this can't be distinguished from a valid [0, 0], so again this is just for testing.) x-tad-bigger___ Chris Pelkie Scientific Visualization Producer 622 Rhodes Hall, Cornell Theory Center Ithaca, NY 14853 /x-tad-bigger On Jun 24, 2005, at 0:37, Lee Hansoo wrote: Thank Chris for your help. I tried other methods you mentioned but I have same problem unresolved yet, they didn't make any differences. It is so weird, I spilted the series data into pieces for every time series and subsituted the missing values to a certain values and tested. but then it is same, just X vaules are shifting. I am sure but the missing values are doubtful since they are all X values without any missing values in Y. Alternatively I have same data in netCDF with variables such as elevation, temperature and wind velocity(wu, wv : this is same wind field) and tried to make 2-vector field with compute module. In this case I don't have any images in ImageWindow, just black, and even any error messages. The modules I used are like below; --> import (wu) Sequencer (import wu, wv separately) --> compute([wu wv]) --> reduce --> include --> autoglyph --> autocolor --> image --> import (wv) Thanks in advance. - Lee HanSoo Research Center for Disaster Environment, Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto, Japan. 611-0011 Tel: 0774-38-4309 M.B : 080-5301-6649
[opendx-users]timeseries 2-vector field generation
Hi OpenDX users I have wind field(nx=130,ny=130) time series (t=60) in ascii file and imported them with mywind.general without problem. However, I have got a problem when I generated result images. The glyphs in U direction(x-axis) were not originated at their origin points, for example (0,0). The origin points of the glyphs are shifting to other points at every timestep. On the other hand in V direction(y-axis) the glyphs are fine in the origin positions. I don't have any ideaon thisproblem, is there any thing about data structure or something like thatI have to keep in mind? Iused following modules and missing value of my data is -. sequencer -- import -- include -- sample-- autoglyph -- autocolor -- image Thanks for any help in advance. -Lee HanSooResearch Center for Disaster Environment, Disaster Prevention Research Institute,Kyoto University, Gokasho, Uji, Kyoto, Japan. 611-0011Tel: 0774-38-4309M.B : 080-5301-6649
[opendx-users]timeseries 2-vector field generation
Thank Chris for your help. I tried other methods you mentioned but I have same problem unresolved yet, they didn't make any differences. It is so weird, I spilted the series data into pieces for every time series and subsituted the missing values to a certain values and tested.but then it is same, just X vaules are shifting. I am sure but the missing values are doubtfulsince they are all Xvalues without any missing values in Y. Alternatively I have same data in netCDF with variables such as elevation, temperature and wind velocity(wu, wv : this is same wind field) and tried to make 2-vector field with compute module. In this case I don't have any images in ImageWindow,just black,and even any error messages. The modules I usedare like below; -- import (wu) Sequencer(importwu, wv separately) -- compute([wu wv])-- reduce -- include --autoglyph --autocolor-- image -- import (wv) Thanks in advance. -Lee HanSooResearch Center for Disaster Environment, Disaster Prevention Research Institute,Kyoto University, Gokasho, Uji, Kyoto, Japan. 611-0011Tel: 0774-38-4309M.B : 080-5301-6649
