comp.lang.idl-pvwave archive: archive » Re: Resampling data with irregular time base

Home » Public Forums » archive » Re: Resampling data with irregular time base

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: Resampling data with irregular time base [message #15692 is a reply to message #15690]

Sat, 05 June 1999 00:00

steinhh
Messages: 260
Registered: June 1994

Senior Member

In article <7jatua$sau$1@kiosk.rzg.mpg.de>
krieger@ipp.mpg.NOSPAM (Karl Krieger) writes:

> I have data with an irregular time base, which I would like to resample
> in a regular spaced time base. How can I average over all original data
> points in each interval of the new time vector without resorting to a
> FOR loop?
> Currently I am using this horrible kludge:
>
> deltat = newtime[1] - newtime[0]
> FOR n=0, n_elements(newtime)-1 DO BEGIN
> index = where((oldtime GT (newtime[n]-deltat/2.)) AND $
> (oldtime LE (newtime[n]+deltat/2.)), $
> count)
> IF count GT 0 THEN newdata[n] = total(olddata[index]) / count
> ENDFOR
>
> Any idea how to transform this in vectorized IDL code? At the moment I
> see no way apart from writing the function in C and calling it by
> linkimage.

I really can't see any other way than to use a temporary weighting
array with size n_elements(newtime) x n_elements(oldtime),
containing the differences between all points in newtime and all
points in oldtime.

If your data sets are large, this would still be very inefficient,
and this problem is a very good example of why IDL will never be
quite optimal compared to fully compiled languages. I C or
Fortran, you'd never muck about with any temporary arrays at all
in this case, you'd calculate the whole thing directly.

I would strongly suggest going for a linkimage (or DLM) function,
something like

newdata = irreg_raverage(oldtime,olddata,newtime,deltat/2.,deltat/2.)

(irregularly sampled running average) allowing you to specify both
(ordered, but possibly irregular!) new sampling times as well as the
forward/backward time windows to average over.

Regards,

Stein Vidar

Report message to a moderator

[Message index]

		Re: Resampling data with irregular time base By: Martin Schultz on Mon, 07 June 1999 00:00
		Re: Resampling data with irregular time base By: Richard G. French on Sat, 05 June 1999 00:00
		Re: Resampling data with irregular time base By: Richard G. French on Sat, 05 June 1999 00:00
		Re: Resampling data with irregular time base By: steinhh on Sat, 05 June 1999 00:00
		Re: Resampling data with irregular time base By: Struan Gray on Tue, 08 June 1999 00:00

Previous Topic:	Re: CW_FSLIDER woes
Next Topic:	Re: pv-wave?

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Oct 11 02:44:25 PDT 2025

Total time taken to generate the page: 1.99485 seconds