| Author |
 Topic  |
|
|
Jartul
Germany
Posts |
 Posted - 09/19/2005 : 1:34:44 PM
|
Origin Version (Select Help-->About Origin): 7.0
Operating System: XP
Hello everybody,
i have two columns of each about 5000 entries. Now I want to reduce them in a sensful way (the shape of the curve should be conserved) to roughly 1000. My idea was to average 5-10 rows to make a new entry. How the hell can I do that? Is this possible with origin, or do I have to play around with excel?
thx in advance
aleX
Example:
5,1
5,15
5,2 --> 5,15
5,1
5,2
next five... and so on |
|
|
Mike
USA
357 Posts |
Posted - 09/19/2005 : 1:54:20 PM
|
Hi Jartul:
If you want to take the averaging approach, you'll probably want a script. Someone may be able to help you with that. If you simply want to return every fifth data point to a new worksheet, do this:
With your worksheet active, select Analysis:Extract Worksheet Data. In the dialog box that opens, enter this in the text ("if") box:
mod(i-1, 5)==0
Click the Do it button. This should copy every 5th data point to a new worksheet.
Mike
OriginLab |
 |
|
|
Jartul
Germany
Posts |
Posted - 09/19/2005 : 2:03:41 PM
|
Thx alot!!!
I dont need to average, so your proposal is the best and shortest way to make a nice picture out of the data points.
I bookmarked this forum, thx again...
aleX |
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 09/20/2005 : 3:37:43 PM
|
Just as a reference...
A sample script of averaging (every 5 rows, from an active worksheet including A(X) and B(Y) may be like this...
ss=5;
%M=%h;
nn=wks.maxRows;
worksheet -b;
%N=%h;
kk=0;
for(ii=1; ii<nn; ii=ii+ss) {
kk++;
totx=0; toty=0;
for(jj=ii; jj<ii+ss; jj++) {totx=totx+%M_a[jj];toty=toty+%M_b[jj];};
%N_a[kk]=totx/ss;
%N_b[kk]=toty/ss;
}
type Original Points:$(nn), Output Rows:$(kk);
OriginC code will be also simple.
--Hideo Fujii
OriginLab
Edited by - Hideo Fujii on 09/20/2005 4:39:48 PM |
|
|
|
g4danny
USA
Posts |
Posted - 09/23/2005 : 7:24:52 PM
|
Interesting problem, for which I used a slightly different approach.
Assume spreadsheet data In[x,y1,y2,...], the data points are not evenly spaced (dx not constant). Instead of averaging a fixed number of data points I fix the dX. The algorithm then is
Out[X,Y1,Y2,...]=Avg[In[x,y1,y2,...]; (X-dX/2)<=x<=(X+dX/2)]
For each X the procedure searches the whole input for rows, where (X-dX/2)<=x<=(X+dX/2), which can be time-costly, but with a decent PC (P4) or Mac (G4) the whole thing takes a second.
I solved this using LabVIEW. Origin script or C should do it faster than that. The strength of my approach is that it significantly reduces noise in my data.
Daniel
Edited by - g4danny on 09/23/2005 7:26:33 PM
Edited by - g4danny on 09/23/2005 7:28:06 PM |
|
|
| |
Topic |
|
|
|
Reducing Datapoints