comp.lang.idl-pvwave archive: archive » Re: Anyway to avoid this last for loop

Home » Public Forums » archive » Re: Anyway to avoid this last for loop

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

Re: Anyway to avoid this last for loop [message #64579 is a reply to message #64572]

Thu, 08 January 2009 09:31

hldevil
Messages: 6
Registered: September 2008

Junior Member

On 8 Jan., 15:28, Tom McGlynn <t...@milkyway.gsfc.nasa.gov> wrote:
> On Jan 8, 4:59 am, hlde...@gmx.de wrote:
>
>> ;AClength is how many frames are thrown out after one AC Hit ->
>> multiple entries can fall into one bin
>
>> detHist=histogram(data,REVERSE_INDICES=ri, /L64, binSize=AClength)
>> acHist=histogram(ACdeletes, /L64, binSize=AClength)
>
>> ;renormalize to one
>> dI=where(detHist GT 1, cntD)
>> aI=where(acHist GT 1, cntA)
>> IF cntD NE 0 THEN detHist[dI]=1
>> IF cntA NE 0 THEN acHist[aI]=1
>
> Can't these last four lines be replaced by
> detHist = detHist<1
> acHist = acHist<1
>
> The < and > operators are useful though not necessarily intuitive.
>
>
>
>
>
>> ;subtract the two histograms. All detector frames which have
>> corresponding AC frame should now have value 0, if AC frame exists but
>> not detector frame then value is -1. If only detector frame exists
>> (the ones we want) value stays 1!!!
>> diffHist=detHist-acHist
>
>> ;keep the ones with one
>> keep=where(diffHist EQ 1, cnt)
>
>> ;And now my problem. I need the indices of the keep-frames but there
>> can be more than one in a bin so I'm stuck with this loop:
>
>> FOR k=0L, cnt-1 DO BEGIN
>> keepIndices=[keepIndices,ri[ri[keep[k]]:ri[keep[k]+1]-1]]
>> ENDFOR
>
> To do it all at once, I think you can just go back to the data array
> as below:
>
> keepHist = where (diffHist EQ 1, cnt) ; From the original
> if (cnt gt 0) then begin
> keepers = intarr(n_elements(detHist))
> keepers[keepHist] = 1;
> keepIndices = where(keepers[data/ACLENGTH] eq 1)
> endif else begin
> keepIndices = [-1]
> endelse
>
> Haven't tested it, but it seems like it or something like it should
> work reasonably efficiently.
>
> Good luck,
> Tom McGlynn

Seems reasonable. I'll try it and see if I get the same results as
with the for loop.

Thanks,

Steffen

Report message to a moderator

[Message index]

		Re: Anyway to avoid this last for loop By: hldevil on Fri, 09 January 2009 00:46
		Re: Anyway to avoid this last for loop By: hldevil on Thu, 08 January 2009 09:31
		Re: Anyway to avoid this last for loop By: David Fanning on Thu, 08 January 2009 06:40
		Re: Anyway to avoid this last for loop By: Tom McGlynn on Thu, 08 January 2009 06:36
		Re: Anyway to avoid this last for loop By: Tom McGlynn on Thu, 08 January 2009 06:28
		Re: Anyway to avoid this last for loop By: R.Bauer on Thu, 08 January 2009 05:54
		Re: Anyway to avoid this last for loop By: Jeremy Bailin on Fri, 09 January 2009 10:28

Previous Topic:	Compile/Restore on Startup
Next Topic:	How to write an array into a dateset as a band, not a dateset?

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Fri Oct 10 15:23:08 PDT 2025

Total time taken to generate the page: 1.35956 seconds